Mumbai.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "MumbaiPaperFinal.ipynb",
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "tdnUmNfZ0rrC"
      },
      "source": [
        "# Unlock Phase-1"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "XQajDkBBUIZJ",
        "outputId": "fa80928a-41a0-4e17-dc87-152e5dde49dd"
      },
      "source": [
        "%reset -f\n",
        "!pip install faiss-gpu\n",
        "!pip install celluloid\n",
        "\n",
        "import faiss\n",
        "import numpy as np\n",
        "from scipy.stats import norm\n",
        "import time\n",
        "import torch\n",
        "from google.colab import files\n",
        "import random\n",
        "from sklearn.neighbors import NearestNeighbors\n",
        "device = 'cuda'\n",
        "res = faiss.StandardGpuResources()\n",
        "select = 0\n",
        "MonteCarlo = 2\n",
        "# Household positions of each agent\n",
        "Residence = torch.tensor(np.loadtxt('mumbai_agent_location.dat'),dtype=torch.double,device='cuda')\n",
        "# Age of the individual\n",
        "Age = torch.tensor(np.loadtxt('mumbai_Age.dat'),dtype=torch.long,device='cuda')\n",
        "# Household size\n",
        "H_Size = torch.tensor(np.loadtxt('mumbai_Age_House.dat'),dtype=torch.long,device='cuda')\n",
        "# 1-member households\n",
        "H_Size[H_Size==0] = 1\n",
        "# ID of the households of each agent\n",
        "H_Number = torch.tensor(np.loadtxt('mumbai_House_number.dat'),dtype=torch.long,device='cuda')\n",
        "# Location of each household\n",
        "H_location = torch.tensor(np.loadtxt('mumbai_household_location.dat'),dtype=torch.long,device='cuda')\n",
        "# Total Population\n",
        "Population = 206340\n",
        "# Number of Children (0), Adults (1), Seniors (2) and Infections (3)\n",
        "Demography = torch.tensor([round(0.2206*Population),round(0.7212*Population),0,round(0.0004345*Population),685],dtype=torch.long,device='cuda')\n",
        "Demography[2] = Population-torch.sum(Demography[:2]) \n",
        "Demography[3] = 800\n",
        "# Layout Size: XMinimum (0), XMaximum (1), YMinimum (2), YMaximum (3), Grid Resolution (4)\n",
        "Layout = torch.tensor([0,5000,0,2000,1000],dtype=torch.double,device='cuda')\n",
        "# Minimum (0), Maximum (1), Inner (2), Innermost (3) Allowed Speeds\n",
        "VelocityParameters = torch.tensor([-500,500,200,10,1],dtype=torch.long,device='cuda')\n",
        "# Collision Radius (0), Infection Radius (1), OuterCircle (2), InnerCircle (3), Reached Goal (4)\n",
        "Radii = torch.tensor([5,2,570,120,5],dtype=torch.long,device='cuda') # Infection Radius(20) > Collision Radius(5/50) --> no social distancing\n",
        "# Total time (days), Resolution(minutes)\n",
        "Time = torch.tensor([31,15],dtype=torch.long, device='cuda') \n",
        "# TimeSteps in a day (1 step = 15mins) and total\n",
        "TimeStepsDay = torch.div(24*60,Time[1].item())\n",
        "# Total time steps in 31 days\n",
        "TimeStepsTotal = TimeStepsDay*Time[0].long()\n",
        "# Re-assign x and y initial positions\n",
        "# Assign positions within unit radius\n",
        "adjustment = 0.5\n",
        "Residence = H_location[H_Number] + torch.DoubleTensor(Population,2).uniform_(-adjustment,adjustment).cuda()\n",
        "# Check if the newly assigned x co-ordinate is within the boundary\n",
        "Residence[Residence[:,0]>Layout[1],0] = (torch.DoubleTensor(Population).uniform_(Layout[1]-adjustment,Layout[1]).cuda())[Residence[:,0]>Layout[1]]\n",
        "Residence[Residence[:,0]<Layout[0],0] = (torch.DoubleTensor(Population).uniform_(Layout[0],Layout[0]+adjustment).cuda())[Residence[:,0]<Layout[0]]\n",
        "# Check if the newly assigned y co-ordinate is within the boundary\n",
        "Residence[Residence[:,1]>Layout[3],1] = (torch.DoubleTensor(Population).uniform_(Layout[3]-adjustment,Layout[3]).cuda())[Residence[:,1]>Layout[3]]\n",
        "Residence[Residence[:,1]<Layout[2],1] = (torch.DoubleTensor(Population).uniform_(Layout[2],Layout[2]+adjustment).cuda())[Residence[:,1]<Layout[2]]\n",
        "Activities = np.zeros((20,96),dtype=np.float32)\n",
        "# Children\n",
        "# Errands \n",
        "Activities[0,24:36]=0.001*norm.pdf(np.arange(6,9,step=0.25), 7.5, 1) # 6-9\n",
        "Activities[0,36:56]=0.002*norm.pdf(np.arange(9,14,step=0.25), 11.5, 1) # 9-14\n",
        "Activities[0,56:72]=0.003*norm.pdf(np.arange(14,18,step=0.25), 17, 1) # 14-21\n",
        "# Adults\n",
        "# Errands \n",
        "Activities[8,24:36]=0.005*norm.pdf(np.arange(6,9,step=0.25), 7.5, 1) # 6-9\n",
        "Activities[8,36:56]=0.01*norm.pdf(np.arange(9,14,step=0.25), 11.5, 1) # 9-14\n",
        "Activities[8,56:72]=0.02*norm.pdf(np.arange(14,18,step=0.25), 17, 1) # 14-21\n",
        "# Nomadic\n",
        "Activities[12,24:48] = 0.05*norm.pdf(np.arange(6,12,step=0.25), 10, 1) # 6-12\n",
        "Activities[12,48:72] = 0.1*norm.pdf(np.arange(12,18,step=0.25), 16, 1) # 12-21\n",
        "# Nomads go home (18) # Weekday\n",
        "Activities[18,0:24] = np.ones(24)\n",
        "Activities[18,84:96] = np.ones(12)\n",
        "# Nomads go home (19) # Weekend\n",
        "Activities[19,0:96] = np.ones(96)\n",
        "# Children\n",
        "# Market\n",
        "Activities[1,24:48]=0.001*norm.pdf(np.arange(6,12,step=0.25), 10, 1) # 6-12\n",
        "Activities[1,48:72]=0.002*norm.pdf(np.arange(12,18,step=0.25), 16, 1) # 12-21\n",
        "# Adults\n",
        "# Market\n",
        "Activities[9,24:48]=0.005*norm.pdf(np.arange(6,12,step=0.25), 10, 1) # 6-12\n",
        "Activities[9,48:72]=0.01*norm.pdf(np.arange(12,18,step=0.25), 17, 1) # 12-21\n",
        "# Office \n",
        "Activities[10,24:48]= 2*norm.pdf(np.arange(6,12,step=0.25), 9, 1) # 6-12\n",
        "Activities[11,48:72] = 2*norm.pdf(np.arange(12,18,step=0.25), 16, 1) # 12-21\n",
        "\n",
        "\n",
        "# Weekday Activities\n",
        "Limits = torch.tensor([0,0,0,0,0,0,1,100,1,100,0,0,\n",
        "                       0,0,0,0,0,0,0,0,0,0,0,0,\n",
        "                       0,0,0,0,2,100,0,100,0,100,0,0,\n",
        "                       0,100,0,0,0,100,0,100,0,100,0,0,\n",
        "                       15,100,0,0,0,100,0,100,0,100,0,100,\n",
        "                       0,0,0,0,0,0,0,0,0,100,0,100],dtype=torch.long,device='cuda')\n",
        "# Weekend Activities                    \n",
        "Limits_Weekend = torch.tensor([0,0,0,0,0,0,0,0,0,0,0,0,\n",
        "                               0,0,0,0,0,0,0,0,0,0,0,0,\n",
        "                               0,0,0,0,0,0,0,0,0,0,0,0,\n",
        "                               0,0,0,0,0,0,0,0,0,0,0,0,\n",
        "                               0,0,0,0,0,0,0,0,0,0,0,0,\n",
        "                               0,0,0,0,0,0,0,0,0,0,0,0],dtype=torch.long,device='cuda')\n",
        "Location_one = torch.tensor([12,13,14,15,16,17],dtype=torch.long,device='cuda') # Location of i\n",
        "Location_two = torch.tensor([[12,0,0,15,16,0],\n",
        "                             [0,0,0,0,0,0],\n",
        "                             [0,0,14,15,16,0],\n",
        "                             [12,0,14,15,16,0],\n",
        "                            [12,0,14,15,16,17],\n",
        "                             [0,0,0,0,16,17]],dtype=torch.long,device='cuda') # Corresponding locations of j  \n",
        "\n",
        "def goal_assign(x,y,minbias,maxbias,xhalfmark,yhalfmark):\n",
        "  '''\n",
        "  Generate Random Goals\n",
        "  '''\n",
        "  # ordinate\n",
        "  l = torch.zeros((Population),dtype=torch.double,device='cuda')\n",
        "  # abscissa\n",
        "  m = torch.zeros((Population),dtype=torch.double,device='cuda')\n",
        "  temp = x -  torch.FloatTensor(Population).uniform_(minbias,maxbias).cuda()\n",
        "  l[x>xhalfmark] = temp[x>xhalfmark] \n",
        "  temp = y -  torch.FloatTensor(Population).uniform_(minbias,maxbias).cuda()\n",
        "  m[y>yhalfmark] = temp[y>yhalfmark] \n",
        "  temp = x +  torch.FloatTensor(Population).uniform_(minbias,maxbias).cuda()\n",
        "  l[x<xhalfmark] = temp[x<xhalfmark] \n",
        "  temp = y +  torch.FloatTensor(Population).uniform_(minbias,maxbias).cuda()\n",
        "  m[y<yhalfmark] = temp[y<yhalfmark] \n",
        "  return l,m   \n",
        "\n",
        "def generateMatrix(agents,matrixone,limits,Population, H_Number, index,con_one, con_two,Time_Daily,matrixthree,distance,radius):\n",
        "  '''\n",
        "  Generate a conditional contact matrix\n",
        "  '''\n",
        "  # Store end result\n",
        "  matrixtwo = torch.zeros((2,Population,20),dtype=torch.double,device='cuda')\n",
        "  count=0\n",
        "  for i in range(6):\n",
        "    vec_con1 = torch.zeros((Population),device='cuda',dtype=torch.bool)\n",
        "    vec_con1 += (agents==con_one[i])\n",
        "    Time_Daily[vec_con1==True,i]+=1\n",
        "    for j in range(6):\n",
        "      temp = torch.zeros((Population,20),device='cuda',dtype=torch.int32)\n",
        "      vec_con2 = torch.zeros((Population,20),device='cuda',dtype=torch.bool)      \n",
        "      if (con_two[i][j]!=0):\n",
        "        if (count==0):          \n",
        "          vec_con2 = torch.logical_and(H_Number[index]==H_Number[:,None],(agents[index]==con_two[i][j]))\n",
        "        else:\n",
        "          vec_con2 = (agents[index]==con_two[i][j])\n",
        "        temp = vec_con1[:,None]*vec_con2\n",
        "        matrixone[:,count] += torch.sum(temp,1)\n",
        "        #matrixone[:,count] += torch.sum(torch.logical_and(temp,distance<radius),1)\n",
        "        if (count!=0):\n",
        "          matrixone[matrixone[:,count]>limits[:,count],count] = limits[matrixone[:,count]>limits[:,count],count]\n",
        "        #tempone = temp*(1-(torch.cumsum(torch.logical_and(temp,distance<radius),axis=1)>matrixone[:,count][:,None]).long())\n",
        "        tempone = temp*(1-(torch.cumsum(temp,axis=1)>matrixone[:,count][:,None]).long())\n",
        "        matrixtwo[0] += tempone\n",
        "        matrixtwo[1] = temp\n",
        "        if (torch.where(vec_con1==True)[0].size(0)!=0):\n",
        "          matrixthree[i,j] += torch.where(tempone==1)[0].size(0)/torch.where(vec_con1==True)[0].size(0)\n",
        "      count+=1\n",
        "  return matrixtwo[0],matrixtwo[1], matrixone, matrixthree, Time_Daily\n",
        "\n",
        "def nearest_destination(Population,centroids,Residence): \n",
        "  '''\n",
        "  Find the nearest destination: Office/School/Market\n",
        "  '''\n",
        "  distance = torch.zeros((Population,len(centroids)),device='cuda',dtype=torch.double)\n",
        "  for i in range(len(centroids)):\n",
        "    # Check the distance from the centroid of the zone to each household location\n",
        "    distance[:,i] = torch.linalg.norm(Residence-Centroids[i],axis=1)\n",
        "  return torch.argmin(distance,axis=1)\n",
        "# Schools\n",
        "# Co-ordinates of different schools\n",
        "School = torch.tensor([[1000,1500,300,500],\n",
        "                       [2000,2500,300,500],\n",
        "                       [3000,3600,500,1000],\n",
        "                       [3500,4000,1300,1500],\n",
        "                       [600,1000,1100,1500],\n",
        "                       [2400,2800,1600,1800]],dtype=torch.long,device='cuda')\n",
        "# Generate Destinations in different school zones\n",
        "School_Positions = torch.zeros((Population,12),dtype=torch.double,device='cuda')\n",
        "# x-co-ordinate\n",
        "School_Positions[:,0] = torch.DoubleTensor(Population).uniform_(School[0,0],School[0,1])\n",
        "School_Positions[:,2] = torch.DoubleTensor(Population).uniform_(School[1,0],School[1,1])\n",
        "School_Positions[:,4] = torch.DoubleTensor(Population).uniform_(School[2,0],School[2,1])\n",
        "School_Positions[:,6] = torch.DoubleTensor(Population).uniform_(School[3,0],School[3,1])\n",
        "School_Positions[:,8] = torch.DoubleTensor(Population).uniform_(School[4,0],School[4,1])\n",
        "School_Positions[:,10] = torch.DoubleTensor(Population).uniform_(School[5,0],School[5,1])\n",
        "# y-co-ordinate\n",
        "School_Positions[:,1] = torch.DoubleTensor(Population).uniform_(School[0,2],School[0,3])\n",
        "School_Positions[:,3] = torch.DoubleTensor(Population).uniform_(School[1,2],School[1,3])\n",
        "School_Positions[:,5] = torch.DoubleTensor(Population).uniform_(School[2,2],School[2,3])\n",
        "School_Positions[:,7] = torch.DoubleTensor(Population).uniform_(School[3,2],School[3,3])\n",
        "School_Positions[:,9] = torch.DoubleTensor(Population).uniform_(School[4,2],School[4,3])\n",
        "School_Positions[:,11] = torch.DoubleTensor(Population).uniform_(School[5,2],School[5,3])\n",
        "\n",
        "# Offices\n",
        "# Co-ordinates of different offices\n",
        "Office = torch.tensor([[1000,1500,0,300],\n",
        "                      [2000,2500,0,300],\n",
        "                      [3500,4000,1000,1300],\n",
        "                      [2000,3000,500,1000]],dtype=torch.long,device='cuda')\n",
        "# Generate destinations in office zones\n",
        "Office_Positions = torch.zeros((Population,8),dtype=torch.double,device='cuda')\n",
        "# x co-ordinate\n",
        "Office_Positions[:,0] = torch.DoubleTensor(Population).uniform_(Office[0,0],Office[0,1])\n",
        "Office_Positions[:,2] = torch.DoubleTensor(Population).uniform_(Office[1,0],Office[1,1])\n",
        "Office_Positions[:,4] = torch.DoubleTensor(Population).uniform_(Office[2,0],Office[2,1])\n",
        "Office_Positions[:,6] = torch.DoubleTensor(Population).uniform_(Office[3,0],Office[3,1])\n",
        "# y co-ordinate\n",
        "Office_Positions[:,1] = torch.DoubleTensor(Population).uniform_(Office[0,2],Office[0,3])\n",
        "Office_Positions[:,3] = torch.DoubleTensor(Population).uniform_(Office[1,2],Office[1,3])\n",
        "Office_Positions[:,5] = torch.DoubleTensor(Population).uniform_(Office[2,2],Office[2,3])\n",
        "Office_Positions[:,7] = torch.DoubleTensor(Population).uniform_(Office[3,2],Office[3,3])\n",
        "\n",
        "# Markets\n",
        "# Co-ordinates of different markets\n",
        "Market = torch.tensor([[2600,3000,1500,2000],\n",
        "                       [3600,4200,100,400],\n",
        "                       [4600,5000,700,900],\n",
        "                       [2000,2400,1000,1200],\n",
        "                       [0,400,1000,1300],\n",
        "                       [3600,4000,1500,1700],\n",
        "                       [1000,2000,500,1000],\n",
        "                       [300,900,1600,1800]],dtype=torch.long,device='cuda')\n",
        "# Generate destinations in market zones\n",
        "Market_Positions = torch.zeros((Population,16),dtype=torch.double,device='cuda')\n",
        "# x co-ordinate\n",
        "Market_Positions[:,0] = torch.DoubleTensor(Population).uniform_(Market[0,0],Market[0,1])\n",
        "Market_Positions[:,2] = torch.DoubleTensor(Population).uniform_(Market[1,0],Market[1,1])\n",
        "Market_Positions[:,4] = torch.DoubleTensor(Population).uniform_(Market[2,0],Market[2,1])\n",
        "Market_Positions[:,6] = torch.DoubleTensor(Population).uniform_(Market[3,0],Market[3,1])\n",
        "Market_Positions[:,8] = torch.DoubleTensor(Population).uniform_(Market[4,0],Market[4,1])\n",
        "Market_Positions[:,10] = torch.DoubleTensor(Population).uniform_(Market[5,0],Market[5,1])\n",
        "Market_Positions[:,12] = torch.DoubleTensor(Population).uniform_(Market[6,0],Market[6,1])\n",
        "Market_Positions[:,14] = torch.DoubleTensor(Population).uniform_(Market[7,0],Market[7,1])\n",
        "# y co-ordinate\n",
        "Market_Positions[:,1] = torch.DoubleTensor(Population).uniform_(Market[0,2],Market[0,3])\n",
        "Market_Positions[:,3] = torch.DoubleTensor(Population).uniform_(Market[1,2],Market[1,3])\n",
        "Market_Positions[:,5] = torch.DoubleTensor(Population).uniform_(Market[2,2],Market[2,3])\n",
        "Market_Positions[:,7] = torch.DoubleTensor(Population).uniform_(Market[3,2],Market[3,3])\n",
        "Market_Positions[:,9] = torch.DoubleTensor(Population).uniform_(Market[4,2],Market[4,3])\n",
        "Market_Positions[:,11] = torch.DoubleTensor(Population).uniform_(Market[5,2],Market[5,3])\n",
        "Market_Positions[:,13] = torch.DoubleTensor(Population).uniform_(Market[6,2],Market[6,3])\n",
        "Market_Positions[:,15] = torch.DoubleTensor(Population).uniform_(Market[7,2],Market[7,3])\n",
        "\n",
        "# Centroids of each zone\n",
        "Centroids = torch.zeros((18,2),dtype=torch.double,device='cuda')\n",
        "# Centroids of school zones\n",
        "Centroids[:6,0] = School[:,0] + 0.5*School[:,1]\n",
        "Centroids[:6,1] = School[:,2] + 0.5*School[:,3]\n",
        "# Centroids of office zones\n",
        "Centroids[6:10,0] = Office[:,0] + 0.5*Office[:,1]\n",
        "Centroids[6:10,1] = Office[:,2] + 0.5*Office[:,3]\n",
        "# Centroids of market zones\n",
        "Centroids[10:18,0] = Market[:,0] + 0.5*Market[:,1]\n",
        "Centroids[10:18,1] = Market[:,2] + 0.5*Market[:,3]\n",
        "\n",
        "# Find nearest destination\n",
        "Nearest = torch.zeros((Population,3),device='cuda',dtype=torch.long)\n",
        "# Nearest school\n",
        "Nearest[:,0] = nearest_destination(Population,Centroids[:6],Residence)\n",
        "# Nearest office\n",
        "Nearest[:,1] = nearest_destination(Population,Centroids[6:10],Residence)\n",
        "# Nearest market\n",
        "Nearest[:,2] = nearest_destination(Population,Centroids[10:18],Residence) \n",
        "# ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n",
        "# Disease Transmission Parameters\n",
        "# -------------------------------------\n",
        "\n",
        "# ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n",
        "# Disease Transmission Parameters\n",
        "# -------------------------------------\n",
        "\n",
        "# States: SIT/SIN/AIT/AIN/SUT/SUN\n",
        "States = torch.tensor(([0,0,0,0,0,1,2,3],\n",
        "                       [0,0,0,1,0,0,0,0],\n",
        "                       [6,7,6,7,0,0,0,0],\n",
        "                       [9,9,9,9,0,0,0,0],\n",
        "                       [0,0,0,0,0,0,0,0],\n",
        "                       [0,1,2,3,4,5,6,7]),dtype=torch.int,device='cuda')\n",
        "# Disease Transition Probabilities\n",
        "Transition = torch.zeros((3,6,8),dtype=torch.double,device='cuda')\n",
        "undercount = 0\n",
        "# Children \n",
        "Transition[0] = torch.transpose(torch.tensor(([0,0,0.12495,0.00005,0,0.875],\n",
        "                                              [0,0,0.09994,0.00006,undercount,0.9-undercount],\n",
        "                                              [0,0.00998,0.19,0.00002,0,0.8],\n",
        "                                              [0,0.02996,0.12,0.00004,undercount,0.85-undercount],\n",
        "                                              [0.2537,0,0,0,0,0.7463],\n",
        "                                              [0.2537,0,0,0,0,0.7463],\n",
        "                                              [0.2537,0,0,0,0,0.7463],\n",
        "                                              [0.2537,0,0,0,0,0.7463]),dtype=torch.double,device='cuda'),0,1) \n",
        "# Adults\n",
        "Transition[1] = torch.transpose(torch.tensor(([0,0,0.124,0.001,0,0.875],\n",
        "                                              [0,0,0.06,0.04,undercount,0.9-undercount],\n",
        "                                              [0,0.03993,0.16,0.00007,0,0.8],\n",
        "                                              [0,0.0493,0.1,0.0007,undercount,0.85-undercount],\n",
        "                                              [0.5,0,0,0,0,0.5],\n",
        "                                              [0.5,0,0,0,0,0.5],\n",
        "                                              [0.5,0,0,0,0,0.5],\n",
        "                                              [0.5,0,0,0,0,0.5]),dtype=torch.double,device='cuda'),0,1) \n",
        "# Seniors\n",
        "Transition[2] = torch.transpose(torch.tensor(([0,0,0.065,0.06,0,0.875],\n",
        "                                              [0,0,0.02,0.08,undercount,0.9-undercount],\n",
        "                                              [0,0.08,0.1,0.02,0,0.8],\n",
        "                                              [0,0.1,0.01,0.04,undercount,0.85-undercount],\n",
        "                                              [0.5951,0,0,0,0,0.4049],\n",
        "                                              [0.5951,0,0,0,0,0.4049],\n",
        "                                              [0.5951,0,0,0,0,0.4049],\n",
        "                                              [0.5951,0,0,0,0,0.4049]),dtype=torch.double,device='cuda'),0,1)\n",
        "# Storage Parameters\n",
        "# Outermost Loop (Scenario/Final Parameters)\n",
        "# Disease Transmission: S/IT/INT/R/D/Total (7+7, Mean and Variance states) \n",
        "FinalTimeSeries = torch.zeros((65,(TimeStepsTotal.long().item())+1), dtype=torch.float,device='cuda')\n",
        "# Time to infection, death, recovery (Mean - 0,1,2, Variance - 3,4,5)\n",
        "FinalConstants = torch.zeros((20), dtype=torch.float,device='cuda')\n",
        "InteractionsFinal = torch.zeros((2,6,6),dtype=torch.float,device='cuda')\n",
        "# Outermost loop\n",
        "# Second Loop (MonteCarlo Run Parameters)\n",
        "# ResultsTimeSeries (Time Series for MC runs)(0-6)\n",
        "ResultsTMC = torch.zeros((MonteCarlo,29,(TimeStepsTotal.long().item())+1), dtype=torch.float,device='cuda')\n",
        "# Results Constants: Time to Infection: Overall(0), Recovery(1), Death(2)\n",
        "ResultsCMC = torch.zeros((MonteCarlo,10,1), dtype=torch.float,device='cuda')\n",
        "InteractionsMC = torch.zeros((MonteCarlo,2,6,6), dtype=torch.float,device='cuda')\n",
        "for mc in range(MonteCarlo):\n",
        "  # Create N agents;\n",
        "  # Age = 0,     \n",
        "  # x_position = 1, \n",
        "  # y_position = 2, \n",
        "  # x_velocity = 3, \n",
        "  # y_velocity = 4, \n",
        "  # occupation = 5,\n",
        "  # current_task = 7, \n",
        "  # status_of_task = 8, \n",
        "  # goal_x = 9, \n",
        "  # goal_y = 10,\n",
        "  # infection_status = 14, \n",
        "  # infection_state = 15, \n",
        "  # time_of_infectiousness = 16\n",
        "  Agents = torch.zeros((Population,20),dtype=torch.double,device='cuda')\n",
        "  # Assign Age \n",
        "  Agents[:,0] = Age\n",
        "  # Indices of Children\n",
        "  Children = torch.where(Agents[:,0]==0)[0]\n",
        "  # Indices of Adults\n",
        "  Adults = torch.where(Agents[:,0]==1)[0]\n",
        "  # Indices of Seniors\n",
        "  Seniors = torch.where(Agents[:,0]==2)[0]\n",
        "  # Assign x and y Positions\n",
        "  # PositionX (1)\n",
        "  Agents[:,1] = Residence[:,0].detach().clone()\n",
        "  # PositionY (2)\n",
        "  Agents[:,2] = Residence[:,1].detach().clone()\n",
        "  # Assign x and y Velocity Components\n",
        "  # VelocityX (3)\n",
        "  Agents[:,3] = torch.DoubleTensor(Population).uniform_(VelocityParameters[0],VelocityParameters[1])\n",
        "  # VelocityY (4)\n",
        "  Agents[:,4] = torch.DoubleTensor(Population).uniform_(VelocityParameters[0],VelocityParameters[1])\n",
        "  # Social Distance:  \n",
        "  SD = Radii[0]*torch.ones((Population),dtype=torch.long,device='cuda')\n",
        "  # Fraction not following (0: Everyone, 1: No one)\n",
        "  SD[torch.tensor([True,False],device='cuda')[torch.multinomial(torch.tensor([0.9,0.1]),Population,replacement=True)]] = 0\n",
        "  # Occupation (5): \n",
        "  # School-going Children (1)\n",
        "  Agents[Children[:int(0.9*Demography[0])],5] = 1\n",
        "  # School-going Adults (1)\n",
        "  Agents[Adults[:int(0.05*Demography[1])],5] = 1\n",
        "  # School-going Seniors (1)\n",
        "  Agents[Seniors[:int(0.02*Demography[2])],5] = 1\n",
        "  # Stay-at-home adults (2)\n",
        "  Agents[Adults[int(0.05*Demography[1]):int(0.25*Demography[1])],5] = 2\n",
        "  # Office Going Adults (3)\n",
        "  Agents[Adults[int(0.25*Demography[1]):int(0.85*Demography[1])],5] = 3\n",
        "  # Nomadic Adults (4)\n",
        "  Agents[Adults[int(0.85*Demography[1]):Demography[1]],5] = 4\n",
        "  # Stay-at-home Seniors (5)\n",
        "  Agents[Seniors[int(0.02*Demography[2]):int(0.62*Demography[2])],5] = 5 \n",
        "  # Office Going Seniors (6)\n",
        "  Agents[Seniors[int(0.62*Demography[2]):int(0.95*Demography[2])],5] = 6\n",
        "  # Reduce office-goers by fraction\n",
        "  stay = torch.where(torch.logical_or(Agents[:,5]==3,Agents[:,5]==6))[0]\n",
        "  staynumber = stay[:int(0.95*stay.size(0))]\n",
        "  Agents[staynumber[Agents[staynumber,0]==1],5] = 2\n",
        "  Agents[staynumber[Agents[staynumber,0]==2],5] = 5\n",
        "  # Nomadic Seniors (7)\n",
        "  Agents[Seniors[int(0.95*Demography[2]):Demography[2]],5] = 7\n",
        "  # Assign Infection Status: Infected and Tested (1), Uninfect(0), Infected and Not Tested (3)\n",
        "  # 0.5% infections among children, 49% infections among adults, 50.5% among seniors\n",
        "  Agents[Children[:int(0.005*Demography[3])],14] = 3 # Infected \n",
        "  Agents[Adults[int(0.005*Demography[3]):int(0.49*Demography[3])],14] = 3 # Infected \n",
        "  Agents[Seniors[int(0.49*Demography[3]):Demography[3]],14] = 3 # Infected \n",
        "  Agents[torch.where(Agents[:,14]==3)[0][:int(Demography[4])],14] = 1 # Confirmed Infections\n",
        "  # Assign Infection State: SIT=0, SIN=1, AIT=2, AIN=3, SUT=4, SUN=5, AUT=6, AUN=7\n",
        "  # Infected\n",
        "  # 80% of the infections are asymptomatic\n",
        "  # 20% of the infections are symptomatic\n",
        "  Agents[torch.where(Agents[:,14]==3)[0][:int(0.8*Demography[3])],15] = 3 # AIN\n",
        "  Agents[torch.where(Agents[:,14]==3)[0][int(0.8*Demography[3]):Demography[3]],15] = 1 # SIN\n",
        "  # Uninfected\n",
        "  # 85% of the uninfected are asymptomatic\n",
        "  # 15% of the uninfected are symptomatic\n",
        "  Agents[torch.where(Agents[:,14]==0)[0][:int(0.85*(Population-Demography[3]))],15] = 7 # AUN\n",
        "  Agents[torch.where(Agents[:,14]==0)[0][int(0.85*(Population-Demography[3])):Population-Demography[3]],15] = 5 # SUN\n",
        "  Agents[torch.where(Agents[:,14]==1)[0],15] = 2\n",
        "  # --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n",
        "  # Placeholder Variables\n",
        "  # -------------------------------------\n",
        "\n",
        "  # Vectors \n",
        "  Vectors = torch.zeros((12,Population),device='cuda',dtype=torch.double)\n",
        "  # Distance Matrix(0), Index Matrix(1), Condition Matrix(2)\n",
        "  Matrixs = torch.zeros((5,Population,20),dtype=torch.double,device='cuda')\n",
        "  # Initial Position \n",
        "  CurrentPosition = torch.dstack((Agents[:,1],Agents[:,2]))\n",
        "  # Initial Velocity\n",
        "  Velocity = torch.dstack((Agents[:,3],Agents[:,4]))\n",
        "  # All agents are initially at home (12)\n",
        "  Agents[:,7] = 12\n",
        "  # Status of the task is complete and initially at home (12)\n",
        "  Agents[:,8] = Agents[:,7].detach().clone()\n",
        "  # Ordinate of goal position\n",
        "  Agents[:,9] = Agents[:,1].detach().clone()\n",
        "  # Abscissa of goal position\n",
        "  Agents[:,10] = Agents[:,2].detach().clone()\n",
        "  # Initial Goals are to remain at home\n",
        "  Destination = torch.dstack((Agents[:,1],Agents[:,2]))\n",
        "  # Days \n",
        "  Days = torch.tensor([0],device='cuda')       \n",
        "  # Distance\n",
        "  Distance = torch.zeros((Population),dtype=torch.double,device='cuda')\n",
        "  # Infection Status over Time\n",
        "  # Overall\n",
        "  Results = torch.tensor([torch.where(Agents[:,14]==0)[0].size()[0], #0:Uninfected\n",
        "                          torch.where(Agents[:,14]==1)[0].size()[0], #1:Infected and Tested\n",
        "                          torch.where(Agents[:,14]==2)[0].size()[0], #2:Unreported Recoveries\n",
        "                          torch.where(Agents[:,14]==3)[0].size()[0], #3:Infected and Not Tested\n",
        "                          torch.where(Agents[:,14]==4)[0].size()[0], #4:Unreported Deaths\n",
        "                          torch.where(Agents[:,14]==5)[0].size()[0], #5:Reported Recoveries\n",
        "                          torch.where(Agents[:,14]==6)[0].size()[0]],dtype=torch.long,device='cuda') #6:Reported Deaths  \n",
        "  # Children\n",
        "  ResultsC = torch.tensor([torch.where(Agents[Children,14]==0)[0].size()[0],\n",
        "                          torch.where(Agents[Children,14]==1)[0].size()[0],\n",
        "                          torch.where(Agents[Children,14]==2)[0].size()[0],\n",
        "                          torch.where(Agents[Children,14]==3)[0].size()[0],\n",
        "                          torch.where(Agents[Children,14]==4)[0].size()[0],\n",
        "                          torch.where(Agents[Children,14]==5)[0].size()[0],\n",
        "                          torch.where(Agents[Children,14]==6)[0].size()[0]],dtype=torch.long,device='cuda')   \n",
        "  # Adults  \n",
        "  ResultsA = torch.tensor([torch.where(Agents[Adults,14]==0)[0].size()[0],\n",
        "                          torch.where(Agents[Adults,14]==1)[0].size()[0],\n",
        "                          torch.where(Agents[Adults,14]==2)[0].size()[0],\n",
        "                          torch.where(Agents[Adults,14]==3)[0].size()[0],\n",
        "                          torch.where(Agents[Adults,14]==4)[0].size()[0],\n",
        "                          torch.where(Agents[Adults,14]==5)[0].size()[0],\n",
        "                          torch.where(Agents[Adults,14]==6)[0].size()[0]],dtype=torch.long,device='cuda')   \n",
        "  # Seniors\n",
        "  ResultsS = torch.tensor([torch.where(Agents[Seniors,14]==0)[0].size()[0],\n",
        "                          torch.where(Agents[Seniors,14]==1)[0].size()[0],\n",
        "                          torch.where(Agents[Seniors,14]==2)[0].size()[0],\n",
        "                          torch.where(Agents[Seniors,14]==3)[0].size()[0],\n",
        "                          torch.where(Agents[Seniors,14]==4)[0].size()[0],\n",
        "                          torch.where(Agents[Seniors,14]==5)[0].size()[0],\n",
        "                          torch.where(Agents[Seniors,14]==6)[0].size()[0]],dtype=torch.long,device='cuda')   \n",
        "\n",
        "  # Interactions: Weekday=0, Weekend=1, Contact_Matrix=2\n",
        "  Interactions = torch.zeros((3,6,6),dtype=torch.double,device='cuda')\n",
        "  # Weekday = 0, Weekend = 1\n",
        "  Time_avg = torch.zeros((2,6),dtype=torch.double,device='cuda')\n",
        "  Time_Daily = torch.zeros((Population,6),dtype=torch.long,device='cuda')\n",
        "  # Number of contacts on a day to day basis: Daily_Contact=0, Contact_Limit=1\n",
        "  Contacts = torch.zeros((2,Population,36),dtype=torch.long,device='cuda')\n",
        "\n",
        "  for t in range(TimeStepsTotal.long().item()):\n",
        "\n",
        "    Days = torch.floor(torch.div(t,TimeStepsDay).cuda())\n",
        "    # Time of the day\n",
        "    time_of_day = t%96 \n",
        "    # Day of the week\n",
        "    day_of_week = Days%7 \n",
        "\n",
        "    # ------------------------------------------------\n",
        "    # Store parameters from the previous run\n",
        "    # ------------------------------------------------\n",
        "\n",
        "    # Infection Status from previous iteration\n",
        "    Vectors[4] = Agents[:,14].detach().clone()\n",
        "    # Infection State from previous iteration\n",
        "    Vectors[5] = Agents[:,15].detach().clone()\n",
        "    # Rtime\n",
        "    Vectors[6] = Vectors[1].detach().clone()\n",
        "\n",
        "    # ---------------------------------------------\n",
        "    # Find Limits for the day\n",
        "    # ---------------------------------------------\n",
        "\n",
        "    if time_of_day==0:\n",
        "      # Assign contact limits for the weekend\n",
        "      if torch.logical_or(day_of_week==5,day_of_week==6):\n",
        "        for i in range(1,36):\n",
        "          Contacts[1][:,i] = Limits_Weekend[2*i+1]\n",
        "      else:\n",
        "        # Assign contact limits for the weekday\n",
        "        for i in range(1,36):\n",
        "          Contacts[1][:,i] = Limits[2*i+1]  \n",
        "\n",
        "    #Agents[:,15]: SIT=0, SIN=1, AIT=2, AIN=3, SUT=4, SUN=5, AUT=6, AUN=7\n",
        "    #Unconfirmed to Confirmed (30 times underreported)\n",
        "    Vectors[0] = torch.rand(Population,device='cuda').double()< 0.000010627#0.00000726 #0.0000450422\n",
        "    Vectors[7] = torch.logical_and(Agents[:,15]==1,Vectors[0]==True)\n",
        "    Agents[Vectors[7].bool(),15] = 0 # SIN --> SIT\n",
        "    Vectors[7] = torch.logical_and(Agents[:,15]==3,Vectors[0]==True)\n",
        "    Agents[Vectors[7].bool(),15] = 2 # AIN --> AIT\n",
        "    # Confirmed Cases\n",
        "    # Infected and Tested\n",
        "    Agents[:,14][torch.logical_or(Agents[:,15]==0,Agents[:,15]==2)] = 1\n",
        "    # Unconfirmed\n",
        "    Agents[:,14][torch.logical_or(Agents[:,15]==1,Agents[:,15]==3)] = 3  \n",
        "\n",
        "    # ---------------------------------------------\n",
        "    # State Transition of Infected Agents\n",
        "    # ---------------------------------------------\n",
        "\n",
        "    # Find all agents who have been infected for more than a day\n",
        "    Vectors[8] = torch.logical_and(torch.logical_and(torch.logical_or(Vectors[4]==1,Vectors[4]==3),Agents[:,16]%TimeStepsDay==0),Agents[:,16]>0)\n",
        "    #Vectors[8] = torch.logical_or(Vectors[4]==1,Vectors[4]==3)\n",
        "    # SIT=0: Is the person in state 0 a child\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==0),Agents[:,0]==0)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,0][torch.multinomial(Transition[0][:,0],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # SIT=0: Is the person in state 0 an adult\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==0),Agents[:,0]==1)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,0][torch.multinomial(Transition[1][:,0],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # SIT=0: Is the person in state 0 a senior\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==0),Agents[:,0]==2)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,0][torch.multinomial(Transition[2][:,0],Population,replacement=True)][Vectors[7].bool()].double()  \n",
        "    # SIN=1: Is the person in state 1 a child\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==1),Agents[:,0]==0)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,1][torch.multinomial(Transition[0][:,1],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # SIN=1: Is the person in state 1 an adult\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==1),Agents[:,0]==1)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,1][torch.multinomial(Transition[1][:,1],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # SIN=1: Is the person in state 1 a senior\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==1),Agents[:,0]==2)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,1][torch.multinomial(Transition[2][:,1],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # AIT=2: Is the person in state 2 a child\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==2),Agents[:,0]==0)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,2][torch.multinomial(Transition[0][:,2],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # AIT=2: Is the person in state 2 an adult\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==2),Agents[:,0]==1)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,2][torch.multinomial(Transition[1][:,2],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # AIT=2: Is the person in state 2 a senior\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==2),Agents[:,0]==2)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,2][torch.multinomial(Transition[2][:,2],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # AIN=3: Is the person in state 3 a child\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==3),Agents[:,0]==0)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,3][torch.multinomial(Transition[0][:,3],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # AIN=3: Is the person in state 3 an adult\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==3),Agents[:,0]==1)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,3][torch.multinomial(Transition[1][:,3],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # AIN=3: Is the person in state 3 a senior\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==3),Agents[:,0]==2)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,3][torch.multinomial(Transition[2][:,3],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # Recovery\n",
        "    Agents[:,14][(Vectors[4]==1) & (Agents[:,15]==6)] = 5 # Reported recoveries\n",
        "    Agents[:,14][(Vectors[4]==3) & (Agents[:,15]==7)] = 2 # Unreported recoveries\n",
        "    # Deaths\n",
        "    Agents[:,14][(Vectors[4]==1) & (Agents[:,15]==9)] = 6 # Reported recoveries\n",
        "    Agents[:,14][(Vectors[4]==3) & (Agents[:,15]==9)] = 4 # Unreported recoveries  \n",
        "    # Time Since Infection\n",
        "    Vectors[8] = torch.logical_or(Agents[:,14]==1,Agents[:,14]==3)\n",
        "    Agents[:,16][Vectors[8].bool()] = Agents[:,16][Vectors[8].bool()]+1\n",
        "    Agents[:,17][Vectors[8].bool()] = torch.round(torch.div(Agents[:,16][Vectors[8].bool()],TimeStepsDay-1)) \n",
        "\n",
        "    # -----------------------------------------\n",
        "    # State Transition of Uninfected Agents\n",
        "    # -----------------------------------------  \n",
        "\n",
        "    # Find agents in the vicinity\n",
        "    CurrentPosition_numpy = np.float32(CurrentPosition[0].cpu().detach().numpy())\n",
        "    # Create a faiss array\n",
        "    index_flat = faiss.IndexFlatL2(CurrentPosition_numpy.shape[1])\n",
        "    gpu_index_flat = faiss.index_cpu_to_gpu(res, 0, index_flat)\n",
        "    gpu_index_flat.add(CurrentPosition_numpy)                  \n",
        "    m, n = gpu_index_flat.search(CurrentPosition_numpy, 20)\n",
        "    Matrixs[1] = torch.from_numpy(n).long()\n",
        "    #Matrixs[1] = torch.reshape(q[q!=torch.arange(0,Population,step=1)[:,None]],(Population,19))\n",
        "    # Distances \n",
        "    Matrixs[0] = torch.squeeze(torch.cdist(CurrentPosition[0][:,None],CurrentPosition[0][Matrixs[1].long()],p=2))\n",
        "    Matrixs[0][Matrixs[1]==torch.arange(0,Population,step=1).cuda()[:,None]] = float(\"Inf\")\n",
        "    # Location Matching and Contact Number Check\n",
        "    Matrixs[2],Matrixs[4],Contacts[0],Interactions[2],Time_Daily = generateMatrix(Agents[:,8],Contacts[0],Contacts[1],Population,H_Number,Matrixs[1].long(),Location_one, Location_two,Time_Daily,Interactions[2],Matrixs[0],Radii[1])\n",
        "\n",
        "    if time_of_day==0:\n",
        "      # Over weekends\n",
        "      if torch.logical_or(day_of_week==5,day_of_week==6):\n",
        "        Time_avg[1] += torch.mean(Time_Daily.double(),0)\n",
        "        Time_Daily = torch.zeros((Population,6),dtype=torch.long,device='cuda')\n",
        "        Interactions[1] += Interactions[2]/48#torch.reshape(torch.sum(Contacts[0],0),(6,6))\n",
        "        Interactions[2] = torch.zeros((6,6),dtype=torch.double,device='cuda')\n",
        "        Contacts[0] = torch.zeros((Population,36),dtype=torch.long,device='cuda')\n",
        "      else:\n",
        "      # Over weekdays\n",
        "        Time_avg[0] += torch.mean(Time_Daily.double(),0)\n",
        "        Time_Daily = torch.zeros((Population,6),dtype=torch.long,device='cuda')\n",
        "        Interactions[0] += Interactions[2]/48#torch.reshape(torch.sum(Contacts[0],0),(6,6))\n",
        "        Interactions[2] = torch.zeros((6,6),dtype=torch.double,device='cuda')\n",
        "        Contacts[0] = torch.zeros((Population,36),dtype=torch.long,device='cuda')\n",
        "  \n",
        "    # Find agents with infections in the vicinity\n",
        "    Vectors[3] = torch.sum(torch.logical_and(torch.logical_and(Matrixs[2].bool()==True,Matrixs[0]<Radii[1]),torch.logical_or(Vectors[4][Matrixs[1].long()]==1,Vectors[4][Matrixs[1].long()]==3)),1)\n",
        "    # Find all uninfected agents with infections in their vicinity\n",
        "    Vectors[8] = torch.logical_and(Vectors[4]==0,Vectors[3].long()>0)\n",
        "    # SUT=4: Is the susceptible person a child\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==4),Agents[:,0]==0)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,4][torch.multinomial(Transition[0][:,4],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # SUT=4: Is the susceptible person an adult\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==4),Agents[:,0]==1)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,4][torch.multinomial(Transition[1][:,4],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # SUT=4: Is the susceptible person a senior\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==4),Agents[:,0]==2)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,4][torch.multinomial(Transition[2][:,4],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # SUN=5: Is the susceptible person a child\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==5),Agents[:,0]==0)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,5][torch.multinomial(Transition[0][:,5],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # SUN=5: Is the susceptible person an adult\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==5),Agents[:,0]==1)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,5][torch.multinomial(Transition[1][:,5],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # SUN=5: Is the susceptible person a senior\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==5),Agents[:,0]==2)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,5][torch.multinomial(Transition[2][:,5],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # AUT=6: Is the susceptible person a child\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==6),Agents[:,0]==0)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,6][torch.multinomial(Transition[0][:,6],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # AUT=6: Is the susceptible person an adult\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==6),Agents[:,0]==1)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,6][torch.multinomial(Transition[1][:,6],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # AUT=6: Is the susceptible person a senior\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==6),Agents[:,0]==2)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,6][torch.multinomial(Transition[2][:,6],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # AUN=7: Is the susceptible person a child\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==7),Agents[:,0]==0)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,7][torch.multinomial(Transition[0][:,7],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # AUN=7: Is the susceptible person an adult\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==7),Agents[:,0]==1)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,7][torch.multinomial(Transition[1][:,7],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # AUN=7: Is the susceptible person a senior\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==7),Agents[:,0]==2)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,7][torch.multinomial(Transition[2][:,7],Population,replacement=True)][Vectors[7].bool()].double() \n",
        "\n",
        "\n",
        "    # -----------------------------------------\n",
        "    # Task Allocation\n",
        "    # ----------------------------------------- \n",
        "\n",
        "    # Weekday\n",
        "    if day_of_week<5:\n",
        "      \n",
        "      # Children\n",
        "      # School going children at home\n",
        "      Vectors[8] = torch.logical_and(torch.logical_and(Agents[:,0]==0,Agents[:,5]==1), Agents[:,8]==12)  \n",
        "      # Destination for school going children: School (1), Random (4), Market (7), Stay at home (12) \n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([1,4,7,12],device='cuda')[torch.multinomial(torch.tensor([Activities[2,time_of_day],Activities[0,time_of_day],Activities[1,time_of_day],1-Activities[0,time_of_day]-Activities[1,time_of_day]-Activities[2,time_of_day]]),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Stay-at-home kids\n",
        "      Vectors[8] = torch.logical_and(Agents[:,5]==0, Agents[:,8]==12)\n",
        "      # Destination for children: Random (4), Market (7), Stay at home (12) \n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([4,7,12],device='cuda')[torch.multinomial(torch.tensor([Activities[0,time_of_day],Activities[1,time_of_day],1-Activities[0,time_of_day]-Activities[1,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Kids at school\n",
        "      Vectors[8] = Agents[:,8]==13\n",
        "      # Remain in School: Stay at school (13), Go home (0) \n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,13],device='cuda')[torch.multinomial(torch.tensor([Activities[3,time_of_day],1-Activities[3,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Kids running errands\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==0,Agents[:,8]==15)\n",
        "      # Stay there (15), Go home (0)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,15],device='cuda')[torch.multinomial(torch.tensor([0.7,0.3],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Kids at the market\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==0,Agents[:,8]==17)\n",
        "      # Stay there (17), Go home (0)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,17],device='cuda')[torch.multinomial(torch.tensor([0.7,0.3],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      \n",
        "      # Adults\n",
        "      # School going adults at home\n",
        "      Vectors[8] = torch.logical_and(torch.logical_and(Agents[:,0]==1,Agents[:,5]==1), Agents[:,8]==12)  \n",
        "      # Destination for school going adults: School (1), Random (5), Market (8), Stay at home (12) \n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([1,5,8,12],device='cuda')[torch.multinomial(torch.tensor([Activities[2,time_of_day],Activities[8,time_of_day],Activities[9,time_of_day],1-Activities[2,time_of_day]-Activities[8,time_of_day]-Activities[9,time_of_day]]),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Office going adults\n",
        "      Vectors[8] = torch.logical_and(Agents[:,5]==3, Agents[:,8]==12)\n",
        "      # Destination for office going adults: Office (2), Random (5), Market (8), Stay home (12)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([2,5,8,12],device='cuda')[torch.multinomial(torch.tensor([Activities[10,time_of_day],Activities[8,time_of_day],Activities[9,time_of_day],1-Activities[8,time_of_day]-Activities[9,time_of_day]-Activities[10,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Stay at home adults\n",
        "      Vectors[8] = torch.logical_and(Agents[:,5]==2,Agents[:,8]==12)  \n",
        "      # Destination for adults: Random (5), Market (8), Stay home (12)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([5,8,12],device='cuda')[torch.multinomial(torch.tensor([Activities[8,time_of_day],Activities[9,time_of_day],1-Activities[8,time_of_day]-Activities[9,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Nomadic adults at home\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==1,torch.logical_and(Agents[:,5]==4,Agents[:,8]==12))\n",
        "      # Adults who are nomads: Go elsewhere (10), Stay home (12)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([10,12],device='cuda')[torch.multinomial(torch.tensor([Activities[12,time_of_day],1-Activities[12,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Nomadic adults on the move\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==1,torch.logical_and(Agents[:,5]==4,Agents[:,8]==16)) \n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,10,16],device='cuda')[torch.multinomial(torch.tensor([Activities[18,time_of_day],Activities[12,time_of_day],1-Activities[18,time_of_day]-Activities[12,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Adults at the office\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==1,Agents[:,8]==14)  \n",
        "      # Stay there (14), Go home (0)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,14],device='cuda')[torch.multinomial(torch.tensor([Activities[11,time_of_day],1-Activities[11,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Adults running errands\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==1,Agents[:,8]==15)  \n",
        "      # Stay there (15), Go home (0)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,15],device='cuda')[torch.multinomial(torch.tensor([0.7,0.3],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Adults at the market\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==1,Agents[:,8]==17)  \n",
        "      # Stay there (17), Go home (0)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,17],device='cuda')[torch.multinomial(torch.tensor([0.7,0.3],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      \n",
        "      # Seniors\n",
        "      # School going seniors at home\n",
        "      Vectors[8] = torch.logical_and(torch.logical_and(Agents[:,0]==2,Agents[:,5]==1), Agents[:,8]==12)  \n",
        "      # Destination for school going seniors: School (1), Random (6), Market (9), Stay at home (12) \n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([1,6,9,12],device='cuda')[torch.multinomial(torch.tensor([Activities[2,time_of_day],Activities[8,time_of_day],Activities[9,time_of_day],1-Activities[2,time_of_day]-Activities[8,time_of_day]-Activities[9,time_of_day]]),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Office going seniors\n",
        "      Vectors[8] = torch.logical_and(Agents[:,5]==6, Agents[:,8]==12)  \n",
        "      # Destination for office going seniors: Office (3), Random (6), Market (9), Stay home (12)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([3,6,9,12],device='cuda')[torch.multinomial(torch.tensor([Activities[10,time_of_day],Activities[8,time_of_day],Activities[9,time_of_day],1-Activities[8,time_of_day]-Activities[9,time_of_day]-Activities[10,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Stay-at-home seniors\n",
        "      Vectors[8] = torch.logical_and(Agents[:,5]==5,Agents[:,8]==12)  \n",
        "      # Destination for seniors: Random (6), Market (9), Stay home (12)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([6,9,12],device='cuda')[torch.multinomial(torch.tensor([Activities[8,time_of_day],Activities[9,time_of_day],1-Activities[8,time_of_day]-Activities[9,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Nomadic Seniors at home\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==2,torch.logical_and(Agents[:,5]==7,Agents[:,8]==12))  \n",
        "      # Seniors who are nomads: Go elsewhere (11), Stay home (12)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([11,12],device='cuda')[torch.multinomial(torch.tensor([Activities[12,time_of_day],1-Activities[12,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Nomadic Seniors on the move\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==2,torch.logical_and(Agents[:,5]==7,Agents[:,8]==16))  \n",
        "      # Seniors who are nomads: Go home (0), Go elsewhere (11), Stay there (16)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,11,16],device='cuda')[torch.multinomial(torch.tensor([Activities[18,time_of_day],Activities[12,time_of_day],1-Activities[12,time_of_day]-Activities[18,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Seniors at office \n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==2,Agents[:,8]==14)  \n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,14],device='cuda')[torch.multinomial(torch.tensor([Activities[11,time_of_day],1-Activities[11,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()  \n",
        "      # Seniors running errands\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==2,Agents[:,8]==15)  \n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,15],device='cuda')[torch.multinomial(torch.tensor([0.7,0.3],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Seniors at the market\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==2,Agents[:,8]==17)  \n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,17],device='cuda')[torch.multinomial(torch.tensor([0.7,0.3],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "\n",
        "    else:\n",
        "\n",
        "      # Children\n",
        "      # School going children\n",
        "      Vectors[8] = torch.logical_and(torch.logical_and(Agents[:,0]==0,Agents[:,5]==1), Agents[:,8]==12)  \n",
        "      # Destination for school going children: School (1), Random (4), Market (7), Stay at home (12) \n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([1,4,7,12],device='cuda')[torch.multinomial(torch.tensor([Activities[6,time_of_day],Activities[4,time_of_day],Activities[5,time_of_day],1-Activities[4,time_of_day]-Activities[5,time_of_day]-Activities[6,time_of_day]]),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Stay-at-home kids\n",
        "      Vectors[8] = torch.logical_and(Agents[:,5]==0, Agents[:,8]==12)\n",
        "      # Destination for children: Random (4), Market (7), Stay at home (12) \n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([4,7,12],device='cuda')[torch.multinomial(torch.tensor([Activities[4,time_of_day],Activities[5,time_of_day],1-Activities[4,time_of_day]-Activities[5,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Kids at school\n",
        "      Vectors[8] = Agents[:,8]==13\n",
        "      # Remain in School: Stay at school (13), Go home (0) \n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,13],device='cuda')[torch.multinomial(torch.tensor([Activities[7,time_of_day],1-Activities[7,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Kids running errands\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==0,Agents[:,8]==15)\n",
        "      # Stay there (15), Go home (0)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,15],device='cuda')[torch.multinomial(torch.tensor([0.7,0.3],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Kids at the market\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==0,Agents[:,8]==17)\n",
        "      # Stay there Market (17), Go home (0)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,17],device='cuda')[torch.multinomial(torch.tensor([0.7,0.3],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      \n",
        "      # Adults\n",
        "      # School going adults at home\n",
        "      Vectors[8] = torch.logical_and(torch.logical_and(Agents[:,0]==1,Agents[:,5]==1), Agents[:,8]==12)  \n",
        "      # Destination for school going adults: School (1), Random (5), Market (8), Stay at home (12) \n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([1,5,8,12],device='cuda')[torch.multinomial(torch.tensor([Activities[6,time_of_day],Activities[13,time_of_day],Activities[14,time_of_day],1-Activities[6,time_of_day]-Activities[13,time_of_day]-Activities[14,time_of_day]]),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Office going adults \n",
        "      Vectors[8] = torch.logical_and(Agents[:,5]==3, Agents[:,8]==12)  \n",
        "      # Destination for office going adults: Office (2), Random (5), Market (8), Stay home (12)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([2,5,8,12],device='cuda')[torch.multinomial(torch.tensor([Activities[15,time_of_day],Activities[13,time_of_day],Activities[14,time_of_day],1-Activities[13,time_of_day]-Activities[14,time_of_day]-Activities[15,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Stay-at-home adults\n",
        "      Vectors[8] = torch.logical_and(Agents[:,5]==2,Agents[:,8]==12)  \n",
        "      # Destination for adults: Random (5), Market (8), Stay home (12)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([5,8,12],device='cuda')[torch.multinomial(torch.tensor([Activities[13,time_of_day],Activities[14,time_of_day],1-Activities[13,time_of_day]-Activities[14,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Nomadic adults at home\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==1,torch.logical_and(Agents[:,5]==4,Agents[:,8]==12))\n",
        "      # Adults who are nomads: Go elsewhere (10), Stay home (12)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([10,12],device='cuda')[torch.multinomial(torch.tensor([Activities[17,time_of_day],1-Activities[17,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Nomadic adults on the move\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==1,torch.logical_and(Agents[:,5]==4,Agents[:,8]==16))\n",
        "      # Adults who are nomads: Go home (0), Go elsehwere (10),Stay there (16)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,10,16],device='cuda')[torch.multinomial(torch.tensor([Activities[19,time_of_day],Activities[17,time_of_day],1-Activities[17,time_of_day]-Activities[19,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Adults running errands\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==1,Agents[:,8]==15)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,15],device='cuda')[torch.multinomial(torch.tensor([0.7,0.3],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Adults returning from market\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==1,Agents[:,8]==17)  \n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,17],device='cuda')[torch.multinomial(torch.tensor([0.7,0.3],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Adults returning from office\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==1,Agents[:,8]==14)  \n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,14],device='cuda')[torch.multinomial(torch.tensor([Activities[16,time_of_day],1-Activities[16,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      \n",
        "      # Seniors\n",
        "      # School going seniors at home\n",
        "      Vectors[8] = torch.logical_and(torch.logical_and(Agents[:,0]==2,Agents[:,5]==1), Agents[:,8]==12)  \n",
        "      # Destination for school going adults: School (1), Random (6), Market (9), Stay at home (12) \n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([1,6,9,12],device='cuda')[torch.multinomial(torch.tensor([Activities[6,time_of_day],Activities[13,time_of_day],Activities[14,time_of_day],1-Activities[6,time_of_day]-Activities[13,time_of_day]-Activities[14,time_of_day]]),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Office going seniors\n",
        "      Vectors[8] = torch.logical_and(Agents[:,5]==6, Agents[:,8]==12)  \n",
        "      # Destination for office going seniors: Office (3), Random (6), Market (9), Stay home (12)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([3,6,9,12],device='cuda')[torch.multinomial(torch.tensor([Activities[15,time_of_day],Activities[13,time_of_day],Activities[14,time_of_day],1-Activities[13,time_of_day]-Activities[14,time_of_day]-Activities[15,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Stay-at-home seniors\n",
        "      Vectors[8] = torch.logical_and(Agents[:,5]==5,Agents[:,8]==12)  \n",
        "      # Destination for seniors: Random (6), Market (9), Stay home (12)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([6,9,12],device='cuda')[torch.multinomial(torch.tensor([Activities[13,time_of_day],Activities[14,time_of_day],1-Activities[13,time_of_day]-Activities[14,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Nomadic Seniors at home\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==2,torch.logical_and(Agents[:,5]==7,Agents[:,8]==12))  \n",
        "      # Seniors who are nomads: Go elsewhere (11), Stay home (12)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([11,12],device='cuda')[torch.multinomial(torch.tensor([Activities[17,time_of_day],1-Activities[17,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Nomadic Seniors on the move\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==2,torch.logical_and(Agents[:,5]==7,Agents[:,8]==16))  \n",
        "      # Seniors who are nomads: Go home(0), Go elsewhere (11), Stay there (16)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,11,16],device='cuda')[torch.multinomial(torch.tensor([Activities[19,time_of_day],Activities[17,time_of_day],1-Activities[19,time_of_day]-Activities[17,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Seniors at office \n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==2,Agents[:,8]==14)  \n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,14],device='cuda')[torch.multinomial(torch.tensor([Activities[16,time_of_day],1-Activities[16,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()  \n",
        "      # Seniors running errands\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==2,Agents[:,8]==15)  \n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,15],device='cuda')[torch.multinomial(torch.tensor([0.7,0.3],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Seniors at the market\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==2,Agents[:,8]==17)  \n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,17],device='cuda')[torch.multinomial(torch.tensor([0.7,0.3],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "    \n",
        "    # --------------------------------\n",
        "    # Assign Task-based Goal Positions\n",
        "    # --------------------------------\n",
        "\n",
        "    for i in range(len(School)):\n",
        "      # Goal of School Going Children\n",
        "      Vectors[8] = torch.logical_and(Nearest[:,0]==i,torch.logical_and(Agents[:,8]==12,Agents[:,7]==1))# Stay at home (12) --> Go to School (1)\n",
        "      Agents[:,9][Vectors[8]==1] = School_Positions[Vectors[8]==1,2*i]\n",
        "      Agents[:,10][Vectors[8]==1] = School_Positions[Vectors[8]==1,2*i+1]\n",
        "    for i in range(len(Office)):\n",
        "      # Goal of Office Going Adults and Seniors\n",
        "      Vectors[8] = torch.logical_and(Nearest[:,0]==i,torch.logical_and(Agents[:,8]==12,torch.logical_or(Agents[:,7]==2,Agents[:,7]==3)))# Stay at home --> Office (Adults/Seniors)\n",
        "      Agents[:,9][Vectors[8]==1] = Office_Positions[Vectors[8]==1,2*i]\n",
        "      Agents[:,10][Vectors[8]==1] = Office_Positions[Vectors[8]==1,2*i+1]\n",
        "    for i in range(len(Market)):\n",
        "      # Goal of Market Going Adults/Seniors/Children\n",
        "      Vectors[8] = torch.logical_and(Nearest[:,1]==i,torch.logical_and(Agents[:,8]==12,torch.logical_or(torch.logical_or(Agents[:,7]==7,Agents[:,7]==8),Agents[:,7]==9))) # Home --> Market\n",
        "      temp1 = torch.where(Vectors[8]==1)[0].size(0)\n",
        "      Agents[:,9][Vectors[8]==1] = Market_Positions[torch.randperm(Population)[:int(temp1)],2*i]\n",
        "      Agents[:,10][Vectors[8]==1] = Market_Positions[torch.randperm(Population)[:int(temp1)],2*i+1]\n",
        "    # Goal of Adults running mobile jobs\n",
        "    mobile = goal_assign(CurrentPosition[0][:,0],CurrentPosition[0][:,1],300,800,0.5*Layout[1],0.5*Layout[3]) #300,800\n",
        "    Vectors[8] = torch.logical_and(Agents[:,0]==1,torch.logical_and(torch.logical_or(Agents[:,8]==16,Agents[:,8]==12),Agents[:,7]==10))\n",
        "    temp1 = torch.where(Vectors[8]==1)[0].size(0)\n",
        "    Agents[:,9][Vectors[8]==1] = mobile[0][Vectors[8]==1]\n",
        "    Agents[:,10][Vectors[8]==1] = mobile[1][Vectors[8]==1]\n",
        "    # Goal of Seniors running mobile jobs\n",
        "    mobile = goal_assign(CurrentPosition[0][:,0],CurrentPosition[0][:,1],300,800,0.5*Layout[1],0.5*Layout[3]) #300,800\n",
        "    Vectors[8] = torch.logical_and(Agents[:,0]==2,torch.logical_and(torch.logical_or(Agents[:,8]==16,Agents[:,8]==12),Agents[:,7]==11))\n",
        "    temp1 = torch.where(Vectors[8]==1)[0].size(0)\n",
        "    Agents[:,9][Vectors[8]==1] = mobile[0][Vectors[8]==1]\n",
        "    Agents[:,10][Vectors[8]==1] = mobile[1][Vectors[8]==1]\n",
        "    # Goal of kids running errands\n",
        "    Vectors[8] = torch.logical_and(Agents[:,8]==12,Agents[:,7]==4)\n",
        "    mobile = goal_assign(Agents[:,1].clone(),Agents[:,2].clone(),300,500,0.5*Layout[1],0.5*Layout[3])#300,500\n",
        "    Agents[:,9][Vectors[8]==1] = mobile[0][Vectors[8]==1]\n",
        "    Agents[:,10][Vectors[8]==1] = mobile[1][Vectors[8]==1]\n",
        "    # Goal of adults running errands\n",
        "    Vectors[8] = torch.logical_and(Agents[:,8]==12,Agents[:,7]==5)\n",
        "    mobile = goal_assign(Agents[:,1].clone(),Agents[:,2].clone(),300,800,0.5*Layout[1],0.5*Layout[3])#300,800\n",
        "    Agents[:,9][Vectors[8]==1] = mobile[0][Vectors[8]==1]\n",
        "    Agents[:,10][Vectors[8]==1] = mobile[1][Vectors[8]==1]\n",
        "    # Goal of seniors running errands\n",
        "    Vectors[8] = torch.logical_and(Agents[:,8]==12,Agents[:,7]==6)\n",
        "    mobile = goal_assign(Agents[:,1].clone(),Agents[:,2].clone(),300,800,0.5*Layout[1],0.5*Layout[3])#300,800\n",
        "    Agents[:,9][Vectors[8]==1] = mobile[0][Vectors[8]==1]\n",
        "    Agents[:,10][Vectors[8]==1] = mobile[1][Vectors[8]==1] \n",
        "    # Goal of Agents Going Home\n",
        "    Vectors[8] = torch.logical_or(torch.logical_and(Agents[:,8]==12,Agents[:,7]==12),Agents[:,7]==0)\n",
        "    Agents[:,9][Vectors[8]==1] = Agents[Vectors[8]==1,1].clone().detach()\n",
        "    Agents[:,10][Vectors[8]==1] = Agents[Vectors[8]==1,2].clone().detach()\n",
        "    \n",
        "    # -----------------------------\n",
        "    # Move Towards a Goal\n",
        "    # -----------------------------\n",
        "\n",
        "    # New goal position corresponding to the assigned task\n",
        "    Destination = torch.dstack((Agents[:,9],Agents[:,10]))\n",
        "    # Distance between the current position and assigned goal\n",
        "    Distance = torch.linalg.norm(CurrentPosition[0]-Destination[0],dim=1)\n",
        "    # Goal following velocity vector\n",
        "    Velocity[0] += Destination[0]-CurrentPosition[0]\n",
        "\n",
        "    # ------------------------------\n",
        "    # Collision Avoidance \n",
        "    # ------------------------------\n",
        "\n",
        "    # Reset avoidance vector in every iteration\n",
        "    VRule = torch.zeros((Population,2),dtype=torch.double,device='cuda') \n",
        "    # Find agents inside critical radius \n",
        "    Matrixs[3] = torch.logical_and(torch.logical_and(Matrixs[0]<SD[:,None],Matrixs[0]!=0),Matrixs[4]==True)\n",
        "    # Collision Avoidance Vector\n",
        "    Difference = CurrentPosition[0][:,None]-CurrentPosition[0][Matrixs[1].long()]\n",
        "    Difference[~Matrixs[3].bool()] = 0\n",
        "    # Number of agents inside the critical radius\n",
        "    Vectors[9] = torch.sum(Matrixs[3],1)\n",
        "    # Normalizing the displacement vector\n",
        "    Difference[Matrixs[3].bool()]/=torch.square(Matrixs[0][Matrixs[3].bool()])[:,None]\n",
        "    VRule[Vectors[9]>0] = torch.div(torch.sum(Difference,1)[Vectors[9]>0],torch.sum(Matrixs[3],1)[Vectors[9]>0][:,None]).double()\n",
        "    # Total Collision avoidance vector component\n",
        "    VRule[Vectors[9]>0] -= Velocity[0][Vectors[9]>0].clone().detach()\n",
        "\n",
        "    # --------------------\n",
        "    # Update Velocity\n",
        "    # --------------------\n",
        "\n",
        "    # Update the velocity of agents\n",
        "    Vectors[8] = torch.logical_and(torch.sum(Matrixs[3],1)>0,Agents[:,8]>=12)\n",
        "    Velocity[0][Vectors[8]==1] += VRule[Vectors[8]==1].clone().detach()\n",
        "\n",
        "    # ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n",
        "    # Limit Speed  \n",
        "    # --------------------  \n",
        "\n",
        "    # Speed is MaximumVelocity when outside the outer circle\n",
        "    Velocity[0][Distance>Radii[2]] = torch.div(Velocity[0][Distance>Radii[2]],torch.norm(Velocity[0],dim=1)[Distance>Radii[2]][:,None])*VelocityParameters[1]   \n",
        "    Vectors[8] = torch.logical_and(Distance<=Radii[2],torch.linalg.norm(Velocity[0],dim=1)!=0)\n",
        "    # Adjustment of velocity according to an exponential function in the intermediate region\n",
        "    Velocity[0][Vectors[8]==1] = torch.div(Velocity[0][Vectors[8]==1],torch.norm(Velocity[0],dim=1)[Vectors[8]==1][:,None])*(1000*torch.exp(-0.009*(Radii[2]-Distance[Vectors[8]==1][:,None])))\n",
        "    Vectors[8] = torch.logical_and(Distance<=Radii[4],torch.linalg.norm(Velocity[0],dim=1)!=0)\n",
        "    # Adjustment of velocity in the innermost circle\n",
        "    Velocity[0][Vectors[8]==1] = torch.div(Velocity[0][Vectors[8]==1],torch.norm(Velocity[0],dim=1)[Vectors[8]==1][:,None])*VelocityParameters[4]\n",
        "    Vectors[8] = torch.logical_and(Distance<=Radii[4],torch.linalg.norm(Velocity[0],dim=1)==0)\n",
        "    # Minimum velocity in the innermost circle\n",
        "    Velocity[0][Vectors[8]==1] = torch.sign(Velocity[0][Vectors[8]==1])*torch.tensor([0.4,0.4],dtype=torch.double,device='cuda')\n",
        "\n",
        "    # --------------------\n",
        "    # Update Position\n",
        "    # --------------------\n",
        "\n",
        "    # Update the position of agents\n",
        "    CurrentPosition[0] += Velocity[0]\n",
        "\n",
        "    # ---------------\n",
        "    # Retention\n",
        "    # ---------------\n",
        "    \n",
        "    # Change position if an agent goes out of the box (Wrap-around)\n",
        "    CurrentPosition[0][CurrentPosition[0][:,0] > Layout[1],0] = torch.DoubleTensor(1).uniform_(Layout[1]-10,Layout[1]).cuda()\n",
        "    CurrentPosition[0][CurrentPosition[0][:,0] < Layout[0],0] = torch.DoubleTensor(1).uniform_(Layout[0],Layout[0]+10).cuda()\n",
        "    CurrentPosition[0][CurrentPosition[0][:,1] > Layout[3],1] = torch.DoubleTensor(1).uniform_(Layout[3]-10,Layout[3]).cuda()\n",
        "    CurrentPosition[0][CurrentPosition[0][:,1] < Layout[2],1] = torch.DoubleTensor(1).uniform_(Layout[2],Layout[2]+10).cuda()\n",
        "\n",
        "    # -------------------\n",
        "    # Current Task Status\n",
        "    # -------------------\n",
        "\n",
        "    # Set status of those within 10 units of their destination as reached, rest in process (9)\n",
        "    # Find distance between current location and goal\n",
        "    Vectors[10] = torch.linalg.norm(CurrentPosition[0]-Destination[0],dim=1)\n",
        "    # If distance is greater than 10, agent is moving towards the goal\n",
        "    Agents[torch.logical_and(Vectors[10]>=Radii[4],Agents[:,7]<=11),8] = Agents[torch.logical_and(Vectors[10]>=Radii[4],Agents[:,7]<=11),7]\n",
        "    # Reached goal\n",
        "    # At home\n",
        "    Agents[torch.logical_or(Agents[:,7]==12,torch.logical_and(Vectors[10]<Radii[4],Agents[:,7]==0)),8] = 12\n",
        "    # At school\n",
        "    Agents[torch.logical_or(Agents[:,7]==13,torch.logical_and(Vectors[10]<Radii[4],Agents[:,7]==1)),8] = 13\n",
        "    # Adults in office\n",
        "    Agents[torch.logical_and(Agents[:,0]==1,torch.logical_or(Agents[:,7]==14,torch.logical_and(Vectors[10]<Radii[4],Agents[:,7]==2))),8] = 14\n",
        "    # Seniors in office\n",
        "    Agents[torch.logical_and(Agents[:,0]==2,torch.logical_or(Agents[:,7]==14,torch.logical_and(Vectors[10]<Radii[4],Agents[:,7]==3))),8] = 14\n",
        "    # Kids running errands\n",
        "    Agents[torch.logical_and(Agents[:,0]==0,torch.logical_or(Agents[:,7]==15,torch.logical_and(Vectors[10]<Radii[4],Agents[:,7]==4))),8] = 15\n",
        "    # Adults running errands\n",
        "    Agents[torch.logical_and(Agents[:,0]==1,torch.logical_or(Agents[:,7]==15,torch.logical_and(Vectors[10]<Radii[4],Agents[:,7]==5))),8] = 15\n",
        "    # Seniors running errands\n",
        "    Agents[torch.logical_and(Agents[:,0]==2,torch.logical_or(Agents[:,7]==15,torch.logical_and(Vectors[10]<Radii[4],Agents[:,7]==6))),8] = 15\n",
        "    # Nomadic adults\n",
        "    Agents[torch.logical_and(Agents[:,0]==1,torch.logical_or(Agents[:,7]==16,torch.logical_and(Vectors[10]<Radii[4],Agents[:,7]==10))),8] = 16\n",
        "    # Nomadic seniors\n",
        "    Agents[torch.logical_and(Agents[:,0]==2,torch.logical_or(Agents[:,7]==16,torch.logical_and(Vectors[10]<Radii[4],Agents[:,7]==11))),8] = 16\n",
        "    # Kids at market\n",
        "    Agents[torch.logical_and(Agents[:,0]==0,torch.logical_or(Agents[:,7]==17,torch.logical_and(Vectors[10]<Radii[4],Agents[:,7]==7))),8] = 17\n",
        "    # Adults at market\n",
        "    Agents[torch.logical_and(Agents[:,0]==1,torch.logical_or(Agents[:,7]==17,torch.logical_and(Vectors[10]<Radii[4],Agents[:,7]==8))),8] = 17\n",
        "    # Seniors at market\n",
        "    Agents[torch.logical_and(Agents[:,0]==2,torch.logical_or(Agents[:,7]==17,torch.logical_and(Vectors[10]<Radii[4],Agents[:,7]==9))),8] = 17\n",
        "\n",
        "    # ---------------\n",
        "    # Compute Time \n",
        "    # --------------\n",
        "\n",
        "    # Time to infection\n",
        "    Vectors[1][Vectors[4]==0] += 1\n",
        "\n",
        "    # ----------------------------\n",
        "    # Compute Reproduction Number\n",
        "    # ----------------------------\n",
        "    # With confirmed infections in vicinity\n",
        "    Vectors[8] = torch.logical_and(Vectors[4]==0,Agents[:,14]==3)\n",
        "    Matrixs[0][torch.logical_and(torch.logical_and(Matrixs[2].bool()==True,Matrixs[0]<Radii[1]),torch.logical_or(Vectors[4][Matrixs[1].long()]==1,Vectors[4][Matrixs[1].long()]==3))==False] = float('inf')\n",
        "    if torch.where(Vectors[8]==True)[0].size()[0]>0:\n",
        "      # Find the nearest individual who infected me \n",
        "      temp = torch.min(Matrixs[0][Vectors[8]==True],1)[1]\n",
        "      for l in range(temp.size()[0]):\n",
        "        Vectors[2][(Matrixs[1].long())[temp[l]]]+=1\n",
        "\n",
        "    # ---------------------------------------------\n",
        "    # Store Values of the Current Iteration\n",
        "    # ---------------------------------------------\n",
        "    #PositionTime = torch.dstack((PositionTime,CurrentPosition))\n",
        "\n",
        "    # Store Infection status for the next iteration\n",
        "    Results = torch.dstack((Results,torch.tensor([torch.where(Agents[:,14]==0)[0].size()[0],\n",
        "                                                  torch.where(Agents[:,14]==1)[0].size()[0],\n",
        "                                                  torch.where(Agents[:,14]==2)[0].size()[0],\n",
        "                                                  torch.where(Agents[:,14]==3)[0].size()[0],\n",
        "                                                  torch.where(Agents[:,14]==4)[0].size()[0],\n",
        "                                                  torch.where(Agents[:,14]==5)[0].size()[0],\n",
        "                                                  torch.where(Agents[:,14]==6)[0].size()[0]],dtype=torch.long,device='cuda')))\n",
        "\n",
        "    ResultsC = torch.dstack((ResultsC,torch.tensor([torch.where(Agents[Children,14]==0)[0].size()[0],\n",
        "                                                    torch.where(Agents[Children,14]==1)[0].size()[0],\n",
        "                                                    torch.where(Agents[Children,14]==2)[0].size()[0],\n",
        "                                                    torch.where(Agents[Children,14]==3)[0].size()[0],\n",
        "                                                    torch.where(Agents[Children,14]==4)[0].size()[0],\n",
        "                                                    torch.where(Agents[Children,14]==5)[0].size()[0],\n",
        "                                                    torch.where(Agents[Children,14]==6)[0].size()[0]],dtype=torch.long,device='cuda')))\n",
        "    \n",
        "    ResultsA = torch.dstack((ResultsA,torch.tensor([torch.where(Agents[Adults,14]==0)[0].size()[0],\n",
        "                                                    torch.where(Agents[Adults,14]==1)[0].size()[0],\n",
        "                                                    torch.where(Agents[Adults,14]==2)[0].size()[0],\n",
        "                                                    torch.where(Agents[Adults,14]==3)[0].size()[0],\n",
        "                                                    torch.where(Agents[Adults,14]==4)[0].size()[0],\n",
        "                                                    torch.where(Agents[Adults,14]==5)[0].size()[0],\n",
        "                                                    torch.where(Agents[Adults,14]==6)[0].size()[0]],dtype=torch.long,device='cuda')))\n",
        "    \n",
        "    ResultsS = torch.dstack((ResultsS,torch.tensor([torch.where(Agents[Seniors,14]==0)[0].size()[0],\n",
        "                                                    torch.where(Agents[Seniors,14]==1)[0].size()[0],\n",
        "                                                    torch.where(Agents[Seniors,14]==2)[0].size()[0],\n",
        "                                                    torch.where(Agents[Seniors,14]==3)[0].size()[0],\n",
        "                                                    torch.where(Agents[Seniors,14]==4)[0].size()[0],\n",
        "                                                    torch.where(Agents[Seniors,14]==5)[0].size()[0],\n",
        "                                                    torch.where(Agents[Seniors,14]==6)[0].size()[0]],dtype=torch.long,device='cuda')))\n",
        "  ResultsTMC[mc][0:7][:] = Results[0] # Disease\n",
        "  ResultsTMC[mc][7:14][:] = ResultsC[0] # Children\n",
        "  ResultsTMC[mc][14:21][:] = ResultsA[0] # Adults\n",
        "  ResultsTMC[mc][21:28][:] = ResultsS[0] # Seniors\n",
        "  # Overall: Time to Infection, Time to Recover, Time to Death\n",
        "  ResultsCMC[mc][0:3] = torch.tensor([[torch.mean(Vectors[1][Vectors[1]>0])/TimeStepsDay.item()],\n",
        "                                    [torch.mean(Agents[:,17][torch.logical_or(Vectors[4]==2,Vectors[4]==5)].double())],\n",
        "                                    [torch.mean(Agents[:,17][torch.logical_or(Vectors[4]==4,Vectors[4]==6)].double())]],device='cuda',dtype=torch.double)\n",
        "  # Overall: Effective Reproduction Number\n",
        "  ResultsCMC[mc][3] = torch.mean(Vectors[2][Vectors[4]!=0])\n",
        "  # Overall: Average time of infectiousness\n",
        "  ResultsCMC[mc][4] = torch.mean(Agents[Vectors[4]!=0,16]/96)\n",
        "\n",
        "  # Overall: Time to Infection, Time to Recover, Time to Death\n",
        "  ResultsCMC[mc][5:8] = torch.tensor([[torch.mean(Vectors[1][Vectors[1]>0])/TimeStepsDay.item()],\n",
        "                                    [torch.mean(Agents[:,17][Vectors[4]==5].double())],\n",
        "                                    [torch.mean(Agents[:,17][Vectors[4]==6].double())]],device='cuda',dtype=torch.double)\n",
        "  # Overall: Effective Reproduction Number\n",
        "  ResultsCMC[mc][8] = torch.mean(Vectors[2][torch.logical_and(Vectors[4]!=0,torch.logical_or(Vectors[4]==1,((Vectors[4]==5)|(Vectors[4]==6))))])\n",
        "  # Overall: Average time of infectiousness\n",
        "  ResultsCMC[mc][9] = torch.mean(Agents[torch.logical_and(Vectors[4]!=0,torch.logical_or(Vectors[4]==1,((Vectors[4]==5)|(Vectors[4]==6)))),16]/96)\n",
        "\n",
        "  # Contact Matrix\n",
        "  # InteractionsMC[mc][0][0] = Interactions[0][0]/Population/23\n",
        "  # InteractionsMC[mc][0][1] = Interactions[0][1]/23/torch.where(Agents[:,5]==1)[0].size(0)\n",
        "  # InteractionsMC[mc][0][2] = Interactions[0][2]/23/torch.where(torch.logical_or(Agents[:,5]==3,Agents[:,5]==6)==True)[0].size(0)\n",
        "  # InteractionsMC[mc][0][3] = Interactions[0][3]/Population/23\n",
        "  # InteractionsMC[mc][0][4] = Interactions[0][4]/torch.where(torch.logical_or(Agents[:,5]==4,Agents[:,5]==7)==True)[0].size(0)/23\n",
        "  # InteractionsMC[mc][0][5] = Interactions[0][5]/Population/23\n",
        "  # InteractionsMC[mc][1][0] = Interactions[1][0]/Population/8\n",
        "  # InteractionsMC[mc][1][1] = Interactions[1][1]/8/torch.where(Agents[:,5]==1)[0].size(0)\n",
        "  # InteractionsMC[mc][1][2] = Interactions[1][2]/8/torch.where(torch.logical_or(Agents[:,5]==3,Agents[:,5]==6)==True)[0].size(0)\n",
        "  # InteractionsMC[mc][1][3] = Interactions[1][3]/Population/8\n",
        "  # InteractionsMC[mc][1][4] = Interactions[1][4]/torch.where(torch.logical_or(Agents[:,5]==4,Agents[:,5]==7)==True)[0].size(0)/8\n",
        "  # InteractionsMC[mc][1][5] = Interactions[1][5]/Population/8\n",
        "\n",
        "  InteractionsMC[mc][0][0] = Interactions[0][0]/23\n",
        "  InteractionsMC[mc][0][1] = Interactions[0][1]/23\n",
        "  InteractionsMC[mc][0][2] = Interactions[0][2]/23\n",
        "  InteractionsMC[mc][0][3] = Interactions[0][3]/23\n",
        "  InteractionsMC[mc][0][4] = Interactions[0][4]/23\n",
        "  InteractionsMC[mc][0][5] = Interactions[0][5]/23\n",
        "  InteractionsMC[mc][1][0] = Interactions[1][0]/8\n",
        "  InteractionsMC[mc][1][1] = Interactions[1][1]/8\n",
        "  InteractionsMC[mc][1][2] = Interactions[1][2]/8\n",
        "  InteractionsMC[mc][1][3] = Interactions[1][3]/8\n",
        "  InteractionsMC[mc][1][4] = Interactions[1][4]/8\n",
        "  InteractionsMC[mc][1][5] = Interactions[1][5]/8\n",
        "\n",
        "# Overall: Confirmed and Cumulative=0, Total Infections (T/N)=1,\n",
        "# Confirmed Infections=2, Confirmed Recoveries=3, Confirmed Deaths=4,\n",
        "# Total Recoveries=5, Total Deaths=6, Cumulative (T/N) = 7\n",
        "FinalTimeSeries[0][:] = torch.mean((ResultsTMC[:][:,1]+ResultsTMC[:][:,6]+ResultsTMC[:][:,5]).double(),0)\n",
        "FinalTimeSeries[1][:] = torch.mean((ResultsTMC[:][:,3]+ResultsTMC[:][:,1]).double(),0)\n",
        "FinalTimeSeries[2][:] = torch.mean(ResultsTMC[:][:,1].double(),0)\n",
        "FinalTimeSeries[3][:] = torch.mean(ResultsTMC[:][:,5].double(),0)\n",
        "FinalTimeSeries[4][:] = torch.mean(ResultsTMC[:][:,6].double(),0)\n",
        "FinalTimeSeries[5][:] = torch.mean((ResultsTMC[:][:,5]+ResultsTMC[:][:,2]).double(),0)\n",
        "FinalTimeSeries[6][:] = torch.mean((ResultsTMC[:][:,6]+ResultsTMC[:][:,4]).double(),0)\n",
        "FinalTimeSeries[7][:] = torch.mean((ResultsTMC[:][:,1]+ResultsTMC[:][:,3]+ResultsTMC[:][:,5]+ResultsTMC[:][:,2]+ResultsTMC[:][:,4]+ResultsTMC[:][:,6]).double(),0)\n",
        "FinalTimeSeries[8][:] = torch.std((ResultsTMC[:][:,1]+ResultsTMC[:][:,6]+ResultsTMC[:][:,5]).double(),0)\n",
        "FinalTimeSeries[9][:] = torch.std((ResultsTMC[:][:,1]+ResultsTMC[:][:,3]).double(),0)\n",
        "FinalTimeSeries[10][:] = torch.std(ResultsTMC[:][:,1].double(),0)\n",
        "FinalTimeSeries[11][:] = torch.std(ResultsTMC[:][:,5].double(),0)\n",
        "FinalTimeSeries[12][:] = torch.std(ResultsTMC[:][:,6].double(),0)\n",
        "FinalTimeSeries[13][:] = torch.std((ResultsTMC[:][:,2]+ResultsTMC[:][:,5]).double(),0)\n",
        "FinalTimeSeries[14][:] = torch.std((ResultsTMC[:][:,4]+ResultsTMC[:][:,6]).double(),0)\n",
        "FinalTimeSeries[15][:] = torch.std((ResultsTMC[:][:,1]+ResultsTMC[:][:,3]+ResultsTMC[:][:,2]+ResultsTMC[:][:,5]+ResultsTMC[:][:,4]+ResultsTMC[:][:,6]).double(),0)\n",
        "\n",
        "# Children: Confirmed and Cumulative=16, Total Infections (T/N)=17,\n",
        "# Confirmed Infections=18, Confirmed Recoveries=19, Confirmed Deaths=20,\n",
        "# Total Recoveries=21, Total Deaths=22, Cumulative (T/N) = 23\n",
        "FinalTimeSeries[16][:] = torch.mean((ResultsTMC[:][:,8]+ResultsTMC[:][:,12]+ResultsTMC[:][:,13]).double(),0)\n",
        "FinalTimeSeries[17][:] = torch.mean((ResultsTMC[:][:,8]+ResultsTMC[:][:,10]).double(),0)\n",
        "FinalTimeSeries[18][:] = torch.mean(ResultsTMC[:][:,8].double(),0)\n",
        "FinalTimeSeries[19][:] = torch.mean(ResultsTMC[:][:,12].double(),0)\n",
        "FinalTimeSeries[20][:] = torch.mean(ResultsTMC[:][:,13].double(),0)\n",
        "FinalTimeSeries[21][:] = torch.mean((ResultsTMC[:][:,12]+ResultsTMC[:][:,9]).double(),0)\n",
        "FinalTimeSeries[22][:] = torch.mean((ResultsTMC[:][:,13]+ResultsTMC[:][:,11]).double(),0)\n",
        "FinalTimeSeries[23][:] = torch.mean((ResultsTMC[:][:,8]+ResultsTMC[:][:,9]+ResultsTMC[:][:,10]+ResultsTMC[:][:,11]+ResultsTMC[:][:,12]+ResultsTMC[:][:,13]).double(),0)                                                                                                                              \n",
        "FinalTimeSeries[24][:] = torch.std((ResultsTMC[:][:,8]+ResultsTMC[:][:,12]+ResultsTMC[:][:,13]).double(),0)\n",
        "FinalTimeSeries[25][:] = torch.std((ResultsTMC[:][:,8]+ResultsTMC[:][:,10]).double(),0)\n",
        "FinalTimeSeries[26][:] = torch.std(ResultsTMC[:][:,8].double(),0)\n",
        "FinalTimeSeries[27][:] = torch.std(ResultsTMC[:][:,12].double(),0)\n",
        "FinalTimeSeries[28][:] = torch.std(ResultsTMC[:][:,13].double(),0)\n",
        "FinalTimeSeries[29][:] = torch.std((ResultsTMC[:][:,12]+ResultsTMC[:][:,9]).double(),0)\n",
        "FinalTimeSeries[30][:] = torch.std((ResultsTMC[:][:,13]+ResultsTMC[:][:,11]).double(),0)\n",
        "FinalTimeSeries[31][:] = torch.std((ResultsTMC[:][:,8]+ResultsTMC[:][:,9]+ResultsTMC[:][:,10]+ResultsTMC[:][:,11]+ResultsTMC[:][:,12]+ResultsTMC[:][:,13]).double(),0)\n",
        "\n",
        "# Adults: Confirmed and Cumulative=32, Total Infections (T/N)=33,\n",
        "# Confirmed Infections=34, Confirmed Recoveries=35, Confirmed Deaths=36,\n",
        "# Total Recoveries=37, Total Deaths=38, Cumulative (T/N) = 39\n",
        "FinalTimeSeries[32][:] = torch.mean((ResultsTMC[:][:,15]+ResultsTMC[:][:,19]+ResultsTMC[:][:,20]).double(),0)\n",
        "FinalTimeSeries[33][:] = torch.mean((ResultsTMC[:][:,15]+ResultsTMC[:][:,17]).double(),0)\n",
        "FinalTimeSeries[34][:] = torch.mean(ResultsTMC[:][:,15].double(),0)\n",
        "FinalTimeSeries[35][:] = torch.mean(ResultsTMC[:][:,19].double(),0)\n",
        "FinalTimeSeries[36][:] = torch.mean(ResultsTMC[:][:,20].double(),0)\n",
        "FinalTimeSeries[37][:] = torch.mean((ResultsTMC[:][:,19]+ResultsTMC[:][:,16]).double(),0)\n",
        "FinalTimeSeries[38][:] = torch.mean((ResultsTMC[:][:,20]+ResultsTMC[:][:,18]).double(),0)\n",
        "FinalTimeSeries[39][:] = torch.mean((ResultsTMC[:][:,15]+ResultsTMC[:][:,16]+ResultsTMC[:][:,17]+ResultsTMC[:][:,18]+ResultsTMC[:][:,19]+ResultsTMC[:][:,20]).double(),0)\n",
        "FinalTimeSeries[40][:] = torch.std((ResultsTMC[:][:,15]+ResultsTMC[:][:,19]+ResultsTMC[:][:,20]).double(),0)\n",
        "FinalTimeSeries[41][:] = torch.std((ResultsTMC[:][:,15]+ResultsTMC[:][:,17]).double(),0)\n",
        "FinalTimeSeries[42][:] = torch.std(ResultsTMC[:][:,15].double(),0)\n",
        "FinalTimeSeries[43][:] = torch.std(ResultsTMC[:][:,19].double(),0)\n",
        "FinalTimeSeries[44][:] = torch.std(ResultsTMC[:][:,20].double(),0)\n",
        "FinalTimeSeries[45][:] = torch.std((ResultsTMC[:][:,19]+ResultsTMC[:][:,16]).double(),0)\n",
        "FinalTimeSeries[46][:] = torch.std((ResultsTMC[:][:,20]+ResultsTMC[:][:,18]).double(),0)\n",
        "FinalTimeSeries[47][:] = torch.std((ResultsTMC[:][:,15]+ResultsTMC[:][:,16]+ResultsTMC[:][:,17]+ResultsTMC[:][:,18]+ResultsTMC[:][:,19]+ResultsTMC[:][:,20]).double(),0)\n",
        "\n",
        "# Seniors: Confirmed and Cumulative=48, Total Infections (T/N)=49,\n",
        "# Confirmed Infections=50, Confirmed Recoveries=51, Confirmed Deaths=52,\n",
        "# Total Recoveries=53, Total Deaths=54, Cumulative (T/N) = 55\n",
        "FinalTimeSeries[48][:] = torch.mean((ResultsTMC[:][:,22]+ResultsTMC[:][:,26]+ResultsTMC[:][:,27]).double(),0)\n",
        "FinalTimeSeries[49][:] = torch.mean((ResultsTMC[:][:,22]+ResultsTMC[:][:,24]).double(),0)\n",
        "FinalTimeSeries[50][:] = torch.mean(ResultsTMC[:][:,22].double(),0)\n",
        "FinalTimeSeries[51][:] = torch.mean(ResultsTMC[:][:,26].double(),0)\n",
        "FinalTimeSeries[52][:] = torch.mean(ResultsTMC[:][:,27].double(),0)\n",
        "FinalTimeSeries[53][:] = torch.mean((ResultsTMC[:][:,23]+ResultsTMC[:][:,26]).double(),0)\n",
        "FinalTimeSeries[54][:] = torch.mean((ResultsTMC[:][:,25]+ResultsTMC[:][:,27]).double(),0)\n",
        "FinalTimeSeries[55][:] = torch.mean((ResultsTMC[:][:,22]+ResultsTMC[:][:,23]+ResultsTMC[:][:,24]+ResultsTMC[:][:,25]+ResultsTMC[:][:,26]+ResultsTMC[:][:,27]).double(),0)\n",
        "FinalTimeSeries[56][:] = torch.std((ResultsTMC[:][:,22]+ResultsTMC[:][:,26]+ResultsTMC[:][:,27]).double(),0)\n",
        "FinalTimeSeries[57][:] = torch.std((ResultsTMC[:][:,22]+ResultsTMC[:][:,24]).double(),0)\n",
        "FinalTimeSeries[58][:] = torch.std(ResultsTMC[:][:,22].double(),0)\n",
        "FinalTimeSeries[59][:] = torch.std(ResultsTMC[:][:,26].double(),0)\n",
        "FinalTimeSeries[60][:] = torch.std(ResultsTMC[:][:,27].double(),0)\n",
        "FinalTimeSeries[61][:] = torch.std((ResultsTMC[:][:,23]+ResultsTMC[:][:,26]).double(),0)\n",
        "FinalTimeSeries[62][:] = torch.std((ResultsTMC[:][:,25]+ResultsTMC[:][:,27]).double(),0)\n",
        "FinalTimeSeries[63][:] = torch.std((ResultsTMC[:][:,22]+ResultsTMC[:][:,23]+ResultsTMC[:][:,24]+ResultsTMC[:][:,25]+ResultsTMC[:][:,26]+ResultsTMC[:][:,27]).double(),0)\n",
        "\n",
        "# Overall: Time to Infection=0, Time to Recover=1, Time to Death=2, Rep Number=3, Time of Infectiousness=4\n",
        "# Confirmed: Time to Infection=5, Time to Recover=6, Time to Death=7, Rep Number=8, Time of Infectiousness=9\n",
        "FinalConstants[0] = torch.mean(ResultsCMC[:][:,0][torch.where(torch.logical_or(torch.isnan(ResultsCMC[:][:,0]),ResultsCMC[:][:,0]==0)==False)].double(),0)\n",
        "FinalConstants[1] = torch.mean(ResultsCMC[:][:,1][torch.where(torch.logical_or(torch.isnan(ResultsCMC[:][:,1]),ResultsCMC[:][:,1]==0)==False)].double(),0)\n",
        "FinalConstants[2] = torch.mean(ResultsCMC[:][:,2][torch.where(torch.logical_or(torch.isnan(ResultsCMC[:][:,2]),ResultsCMC[:][:,2]==0)==False)].double(),0)\n",
        "FinalConstants[3] = torch.std(ResultsCMC[:][:,0][torch.where(torch.logical_or(torch.isnan(ResultsCMC[:][:,0]),ResultsCMC[:][:,0]==0)==False)].double(),0)\n",
        "FinalConstants[4] = torch.std(ResultsCMC[:][:,1][torch.where(torch.logical_or(torch.isnan(ResultsCMC[:][:,1]),ResultsCMC[:][:,1]==0)==False)].double(),0)\n",
        "FinalConstants[5] = torch.std(ResultsCMC[:][:,2][torch.where(torch.logical_or(torch.isnan(ResultsCMC[:][:,2]),ResultsCMC[:][:,2]==0)==False)].double(),0) \n",
        "FinalConstants[6] = torch.mean(ResultsCMC[:][:,3].double())\n",
        "FinalConstants[7] = torch.std(ResultsCMC[:][:,3].double())\n",
        "FinalConstants[8] = torch.mean(ResultsCMC[:][:,4].double())\n",
        "FinalConstants[9] = torch.std(ResultsCMC[:][:,4].double())\n",
        "FinalConstants[10] = torch.mean(ResultsCMC[:][:,5][torch.where(torch.logical_or(torch.isnan(ResultsCMC[:][:,5]),ResultsCMC[:][:,5]==0)==False)].double(),0)\n",
        "FinalConstants[11] = torch.mean(ResultsCMC[:][:,6][torch.where(torch.logical_or(torch.isnan(ResultsCMC[:][:,6]),ResultsCMC[:][:,6]==0)==False)].double(),0)\n",
        "FinalConstants[12] = torch.mean(ResultsCMC[:][:,7][torch.where(torch.logical_or(torch.isnan(ResultsCMC[:][:,7]),ResultsCMC[:][:,7]==0)==False)].double(),0)\n",
        "FinalConstants[13] = torch.std(ResultsCMC[:][:,5][torch.where(torch.logical_or(torch.isnan(ResultsCMC[:][:,5]),ResultsCMC[:][:,5]==0)==False)].double(),0)\n",
        "FinalConstants[14] = torch.std(ResultsCMC[:][:,6][torch.where(torch.logical_or(torch.isnan(ResultsCMC[:][:,6]),ResultsCMC[:][:,6]==0)==False)].double(),0)\n",
        "FinalConstants[15] = torch.std(ResultsCMC[:][:,7][torch.where(torch.logical_or(torch.isnan(ResultsCMC[:][:,7]),ResultsCMC[:][:,7]==0)==False)].double(),0) \n",
        "FinalConstants[16] = torch.mean(ResultsCMC[:][:,8].double())\n",
        "FinalConstants[17] = torch.std(ResultsCMC[:][:,8].double())\n",
        "FinalConstants[18] = torch.mean(ResultsCMC[:][:,9].double())\n",
        "FinalConstants[19] = torch.std(ResultsCMC[:][:,9].double())\n",
        "\n",
        "InteractionsFinal[0] = torch.mean(InteractionsMC[:][:,0],0)\n",
        "InteractionsFinal[1] = torch.mean(InteractionsMC[:][:,1],0)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Requirement already satisfied: faiss-gpu in /usr/local/lib/python3.7/dist-packages (1.7.1.post3)\n",
            "Requirement already satisfied: celluloid in /usr/local/lib/python3.7/dist-packages (0.2.0)\n",
            "Requirement already satisfied: matplotlib in /usr/local/lib/python3.7/dist-packages (from celluloid) (3.2.2)\n",
            "Requirement already satisfied: pyparsing!=2.0.4,!=2.1.2,!=2.1.6,>=2.0.1 in /usr/local/lib/python3.7/dist-packages (from matplotlib->celluloid) (3.0.6)\n",
            "Requirement already satisfied: cycler>=0.10 in /usr/local/lib/python3.7/dist-packages (from matplotlib->celluloid) (0.11.0)\n",
            "Requirement already satisfied: python-dateutil>=2.1 in /usr/local/lib/python3.7/dist-packages (from matplotlib->celluloid) (2.8.2)\n",
            "Requirement already satisfied: kiwisolver>=1.0.1 in /usr/local/lib/python3.7/dist-packages (from matplotlib->celluloid) (1.3.2)\n",
            "Requirement already satisfied: numpy>=1.11 in /usr/local/lib/python3.7/dist-packages (from matplotlib->celluloid) (1.19.5)\n",
            "Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.7/dist-packages (from python-dateutil>=2.1->matplotlib->celluloid) (1.15.0)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TNsXt5PezUG3"
      },
      "source": [
        "##Plots"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wbaA753wVRhT"
      },
      "source": [
        "from matplotlib import pyplot as plt\n",
        "from IPython.display import HTML\n",
        "import seaborn as sns\n",
        "from matplotlib.ticker import AutoMinorLocator, MultipleLocator\n",
        "from matplotlib.colors import ListedColormap, LinearSegmentedColormap\n",
        "from matplotlib import collections as matcoll\n",
        "from matplotlib import cm\n",
        "import numpy as np\n",
        "from scipy.stats import norm"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 528
        },
        "id": "7zhKtuU1zb8y",
        "outputId": "94762f01-5cd6-4ddb-f360-fc37d96c0b0c"
      },
      "source": [
        "fig, ax = plt.subplots(figsize=(8, 8))\n",
        "\n",
        "s = 96\n",
        "\n",
        "t=96*31\n",
        "\n",
        "x = (np.arange(0,t,step=s)/96)+1\n",
        "\n",
        "data = [0.332,0.341,0.351,0.363,0.372,0.382,0.394,0.405,0.413,0.425,0.437,0.448,\n",
        "        0.459,0.470,0.479,0.486,0.497,0.508,0.518,0.528,0.537,0.546,0.553,0.562,\n",
        "        0.573,0.583,0.600,0.610,0.620,0.627,0.639]\n",
        "\n",
        "cinf_mean = ((FinalTimeSeries[0]/Population)*100).cpu().detach().numpy()\n",
        "cinf_std = ((FinalTimeSeries[8]/Population)*100).cpu().detach().numpy()\n",
        "\n",
        "ainf_mean = ((FinalTimeSeries[2]/Population)*100).cpu().detach().numpy()\n",
        "ainf_std = ((FinalTimeSeries[10]/Population)*100).cpu().detach().numpy()\n",
        "\n",
        "recovered_mean = ((FinalTimeSeries[3]/Population)*100).cpu().detach().numpy()\n",
        "recovered_std = ((FinalTimeSeries[11]/Population)*100).cpu().detach().numpy()\n",
        "\n",
        "deceased_mean = ((FinalTimeSeries[4]/Population)*100).cpu().detach().numpy()\n",
        "deceased_std = ((FinalTimeSeries[12]/Population)*100).cpu().detach().numpy() \n",
        "\n",
        "rinf_mean = recovered_mean\n",
        "rinf_std = recovered_std\n",
        "\n",
        "dinf_mean = deceased_mean\n",
        "dinf_std = deceased_std\n",
        "\n",
        "plt.plot(x,cinf_mean[np.arange(0,t,step=s)],'x-',color='blue',linewidth=2,linestyle='-')\n",
        "ax.fill_between(x, np.clip(cinf_mean[np.arange(0,t,step=s)]-cinf_std[np.arange(0,t,step=s)],0,np.inf), np.clip(cinf_mean[np.arange(0,t,step=s)]+cinf_std[np.arange(0,t,step=s)],0,np.inf) ,alpha=0.3, facecolor='blue')\n",
        "\n",
        "plt.plot(x,ainf_mean[np.arange(0,t,step=s)],'x-',color='r',linewidth=2,linestyle='-')\n",
        "ax.fill_between(x, np.clip(ainf_mean[np.arange(0,t,step=s)]-ainf_std[np.arange(0,t,step=s)],0,np.inf), np.clip(ainf_mean[np.arange(0,t,step=s)]+ainf_std[np.arange(0,t,step=s)],0,np.inf) ,alpha=0.3, facecolor='r')\n",
        "\n",
        "plt.plot(x,rinf_mean[np.arange(0,t,step=s)],'x-',color='g',linewidth=2,linestyle='-')\n",
        "ax.fill_between(x, np.clip(rinf_mean[np.arange(0,t,step=s)]-rinf_std[np.arange(0,t,step=s)],0,np.inf), np.clip(rinf_mean[np.arange(0,t,step=s)]+rinf_std[np.arange(0,t,step=s)],0,np.inf) ,alpha=0.3, facecolor='g')\n",
        "\n",
        "plt.plot(x,dinf_mean[np.arange(0,t,step=s)],'x-',color='dimgray',linewidth=2,linestyle='-')\n",
        "ax.fill_between(x, np.clip(dinf_mean[np.arange(0,t,step=s)]-dinf_std[np.arange(0,t,step=s)],0,np.inf), np.clip(dinf_mean[np.arange(0,t,step=s)]+dinf_std[np.arange(0,t,step=s)],0,np.inf) ,alpha=0.3, facecolor='dimgray')\n",
        "\n",
        "plt.plot(np.arange(1,32,step=1),data,'o-',color='black',linewidth=2)\n",
        "\n",
        "plt.rcParams['figure.facecolor'] = 'ivory'\n",
        "\n",
        "plt.rcParams[\"font.weight\"] = \"bold\"\n",
        "\n",
        "plt.rcParams[\"axes.labelweight\"] = \"bold\"\n",
        "\n",
        "ax.grid()\n",
        "\n",
        "ax.set_axisbelow(True)\n",
        "ax.xaxis.set_major_locator(plt.MaxNLocator(16))\n",
        "ax.tick_params(which='major', length=7, color='black')\n",
        "ax.tick_params(which='minor', length=4, color='black' )\n",
        "ax.xaxis.set_minor_locator(plt.MaxNLocator())\n",
        "\n",
        "ax.tick_params(which='both', # Options for both major and minor ticks\n",
        "               top=False, # turn off top ticks\n",
        "               left=False, # turn off left ticks\n",
        "               right=False,  # turn off right ticks\n",
        "               bottom=True) # turn off bottom ticks\n",
        "ax.set_xticks(np.arange(1,32,step=2))\n",
        "ax.set_xticklabels([\"1\",\"3\",\"5\",\"7\",\"9\",\"11\",\"13\",\"15\",\"17\",\"19\",\"21\",\"23\",\"25\",\n",
        "                    \"28\",\"29\",\"31\"])\n",
        "\n",
        "plt.xticks(fontsize=16,fontweight='bold')\n",
        "plt.yticks(np.arange(0,1.1,step=0.1),fontsize=16,fontweight='bold')\n",
        "\n",
        "plt.xlabel('Time (Days)',fontsize=16,fontweight='bold')\n",
        "\n",
        "plt.ylabel('Fraction of the total population (%)',fontsize=16,fontweight='bold')\n",
        "\n",
        "plt.legend(['Total Infections','Active Infections','Recoveries','Mortalities','Data'],\n",
        "           fontsize=16,loc=\"upper left\",fancybox=True, \n",
        "           framealpha=0.0, shadow=False, borderpad=1)\n",
        "\n",
        "plt.title('Disease evolution in Mumbai',fontsize=16,fontweight='bold')\n",
        "\n",
        "plt.savefig('FM_1.pdf',format='pdf', bbox_inches='tight',Transparent=True)\n",
        "\n",
        "files.download('FM_1.pdf')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "\n",
              "    async function download(id, filename, size) {\n",
              "      if (!google.colab.kernel.accessAllowed) {\n",
              "        return;\n",
              "      }\n",
              "      const div = document.createElement('div');\n",
              "      const label = document.createElement('label');\n",
              "      label.textContent = `Downloading \"${filename}\": `;\n",
              "      div.appendChild(label);\n",
              "      const progress = document.createElement('progress');\n",
              "      progress.max = size;\n",
              "      div.appendChild(progress);\n",
              "      document.body.appendChild(div);\n",
              "\n",
              "      const buffers = [];\n",
              "      let downloaded = 0;\n",
              "\n",
              "      const channel = await google.colab.kernel.comms.open(id);\n",
              "      // Send a message to notify the kernel that we're ready.\n",
              "      channel.send({})\n",
              "\n",
              "      for await (const message of channel.messages) {\n",
              "        // Send a message to notify the kernel that we're ready.\n",
              "        channel.send({})\n",
              "        if (message.buffers) {\n",
              "          for (const buffer of message.buffers) {\n",
              "            buffers.push(buffer);\n",
              "            downloaded += buffer.byteLength;\n",
              "            progress.value = downloaded;\n",
              "          }\n",
              "        }\n",
              "      }\n",
              "      const blob = new Blob(buffers, {type: 'application/binary'});\n",
              "      const a = document.createElement('a');\n",
              "      a.href = window.URL.createObjectURL(blob);\n",
              "      a.download = filename;\n",
              "      div.appendChild(a);\n",
              "      a.click();\n",
              "      div.remove();\n",
              "    }\n",
              "  "
            ],
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "download(\"download_ddb9d197-def3-49be-9891-c7008c6ff3da\", \"FM_1.pdf\", 21716)"
            ],
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x576 with 1 Axes>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 362
        },
        "id": "QklUZvYqpKWB",
        "outputId": "7b8672cf-4a4f-427b-cab0-8cd1aab9245f"
      },
      "source": [
        "fig,ax = plt.subplots(figsize=(6, 6))\n",
        "\n",
        "data1 = InteractionsFinal[0].cpu().detach().numpy()\n",
        "\n",
        "data1 = np.ma.masked_where(data1 == 0, data1)\n",
        "\n",
        "data2 = InteractionsFinal[0].cpu().detach().numpy()\n",
        "\n",
        "data1[0,0] = 4.5\n",
        "\n",
        "data2[0,0] = 4.5\n",
        "\n",
        "vmin = np.min(np.array([data1,data2]))\n",
        "\n",
        "vmax = np.max(np.array([data1,data2]))\n",
        "\n",
        "alpha = ['$Q_{1}$','$Q_{2}$','$Q_{3}$','$Q_{4}$','$Q_{5}$','$Q_{6}$']\n",
        "\n",
        "cmap = ListedColormap([\"lemonchiffon\",\"goldenrod\",\"cyan\",\"orange\",\"gainsboro\", \"black\",\"yellow\"])\n",
        "\n",
        "cax = ax.matshow(data1,interpolation='nearest',cmap=cmap, rasterized=True, vmin=vmin, vmax=vmax)\n",
        "\n",
        "cbar=fig.colorbar(cax,extend='min', label='Average Contacts')\n",
        "\n",
        "plt.gca().set_xticks([x - 0.5 for x in plt.gca().get_xticks()][1:], minor='true')\n",
        "\n",
        "plt.gca().set_yticks([y - 0.5 for y in plt.gca().get_yticks()][1:], minor='true')\n",
        "\n",
        "plt.grid(which='minor')\n",
        "\n",
        "cmap.set_bad(color='white')\n",
        "\n",
        "ax.set_xticklabels(['']+alpha,y=-0.15,fontsize=12)\n",
        "\n",
        "ax.set_yticklabels(['']+alpha,fontsize=12)\n",
        "\n",
        "cbar.ax.tick_params(labelsize=12)\n",
        "\n",
        "plt.title('Weekday daily-contacts matrix',fontweight='bold',fontsize=14)\n",
        "\n",
        "plt.savefig('IFM_1.pdf',format='pdf', bbox_inches='tight',Transparent=True)\n",
        "\n",
        "files.download('IFM_1.pdf')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "\n",
              "    async function download(id, filename, size) {\n",
              "      if (!google.colab.kernel.accessAllowed) {\n",
              "        return;\n",
              "      }\n",
              "      const div = document.createElement('div');\n",
              "      const label = document.createElement('label');\n",
              "      label.textContent = `Downloading \"${filename}\": `;\n",
              "      div.appendChild(label);\n",
              "      const progress = document.createElement('progress');\n",
              "      progress.max = size;\n",
              "      div.appendChild(progress);\n",
              "      document.body.appendChild(div);\n",
              "\n",
              "      const buffers = [];\n",
              "      let downloaded = 0;\n",
              "\n",
              "      const channel = await google.colab.kernel.comms.open(id);\n",
              "      // Send a message to notify the kernel that we're ready.\n",
              "      channel.send({})\n",
              "\n",
              "      for await (const message of channel.messages) {\n",
              "        // Send a message to notify the kernel that we're ready.\n",
              "        channel.send({})\n",
              "        if (message.buffers) {\n",
              "          for (const buffer of message.buffers) {\n",
              "            buffers.push(buffer);\n",
              "            downloaded += buffer.byteLength;\n",
              "            progress.value = downloaded;\n",
              "          }\n",
              "        }\n",
              "      }\n",
              "      const blob = new Blob(buffers, {type: 'application/binary'});\n",
              "      const a = document.createElement('a');\n",
              "      a.href = window.URL.createObjectURL(blob);\n",
              "      a.download = filename;\n",
              "      div.appendChild(a);\n",
              "      a.click();\n",
              "      div.remove();\n",
              "    }\n",
              "  "
            ],
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "download(\"download_02aae039-e201-4f5e-9e54-c598cb92010f\", \"IFM_1.pdf\", 16441)"
            ],
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x432 with 2 Axes>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 277
        },
        "id": "EtG3hys8pvP3",
        "outputId": "45d1c673-930d-4d04-98f5-5479ca7dd01b"
      },
      "source": [
        "!pip install hickle\n",
        "import hickle as hk\n",
        "from google.colab import files\n",
        "hk.dump(FinalTimeSeries.cpu().detach().numpy(),'TS_FM_1.hkl', mode='w')\n",
        "hk.dump(FinalConstants.cpu().detach().numpy(),'C_FM_1.hkl', mode='w')\n",
        "hk.dump(InteractionsFinal.cpu().detach().numpy(),'I_FM_1.hkl', mode='w')\n",
        "files.download('TS_FM_1.hkl')\n",
        "files.download('C_FM_1.hkl')\n",
        "files.download('I_FM_1.hkl')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Collecting hickle\n",
            "  Downloading hickle-4.0.4-py3-none-any.whl (49 kB)\n",
            "\u001b[?25l\r\u001b[K     |██████▊                         | 10 kB 30.2 MB/s eta 0:00:01\r\u001b[K     |█████████████▍                  | 20 kB 27.6 MB/s eta 0:00:01\r\u001b[K     |████████████████████            | 30 kB 18.3 MB/s eta 0:00:01\r\u001b[K     |██████████████████████████▊     | 40 kB 15.6 MB/s eta 0:00:01\r\u001b[K     |████████████████████████████████| 49 kB 2.8 MB/s \n",
            "\u001b[?25hRequirement already satisfied: dill>=0.3.0 in /usr/local/lib/python3.7/dist-packages (from hickle) (0.3.4)\n",
            "Requirement already satisfied: numpy>=1.8 in /usr/local/lib/python3.7/dist-packages (from hickle) (1.19.5)\n",
            "Requirement already satisfied: six>=1.11.0 in /usr/local/lib/python3.7/dist-packages (from hickle) (1.15.0)\n",
            "Collecting h5py<3.0.0,>=2.8.0\n",
            "  Downloading h5py-2.10.0-cp37-cp37m-manylinux1_x86_64.whl (2.9 MB)\n",
            "\u001b[K     |████████████████████████████████| 2.9 MB 8.5 MB/s \n",
            "\u001b[?25hInstalling collected packages: h5py, hickle\n",
            "  Attempting uninstall: h5py\n",
            "    Found existing installation: h5py 3.1.0\n",
            "    Uninstalling h5py-3.1.0:\n",
            "      Successfully uninstalled h5py-3.1.0\n",
            "Successfully installed h5py-2.10.0 hickle-4.0.4\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "\n",
              "    async function download(id, filename, size) {\n",
              "      if (!google.colab.kernel.accessAllowed) {\n",
              "        return;\n",
              "      }\n",
              "      const div = document.createElement('div');\n",
              "      const label = document.createElement('label');\n",
              "      label.textContent = `Downloading \"${filename}\": `;\n",
              "      div.appendChild(label);\n",
              "      const progress = document.createElement('progress');\n",
              "      progress.max = size;\n",
              "      div.appendChild(progress);\n",
              "      document.body.appendChild(div);\n",
              "\n",
              "      const buffers = [];\n",
              "      let downloaded = 0;\n",
              "\n",
              "      const channel = await google.colab.kernel.comms.open(id);\n",
              "      // Send a message to notify the kernel that we're ready.\n",
              "      channel.send({})\n",
              "\n",
              "      for await (const message of channel.messages) {\n",
              "        // Send a message to notify the kernel that we're ready.\n",
              "        channel.send({})\n",
              "        if (message.buffers) {\n",
              "          for (const buffer of message.buffers) {\n",
              "            buffers.push(buffer);\n",
              "            downloaded += buffer.byteLength;\n",
              "            progress.value = downloaded;\n",
              "          }\n",
              "        }\n",
              "      }\n",
              "      const blob = new Blob(buffers, {type: 'application/binary'});\n",
              "      const a = document.createElement('a');\n",
              "      a.href = window.URL.createObjectURL(blob);\n",
              "      a.download = filename;\n",
              "      div.appendChild(a);\n",
              "      a.click();\n",
              "      div.remove();\n",
              "    }\n",
              "  "
            ],
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "download(\"download_e75d5ca1-b45e-48ad-b7e9-a4f04e4a849a\", \"TS_FM_1.hkl\", 780164)"
            ],
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "\n",
              "    async function download(id, filename, size) {\n",
              "      if (!google.colab.kernel.accessAllowed) {\n",
              "        return;\n",
              "      }\n",
              "      const div = document.createElement('div');\n",
              "      const label = document.createElement('label');\n",
              "      label.textContent = `Downloading \"${filename}\": `;\n",
              "      div.appendChild(label);\n",
              "      const progress = document.createElement('progress');\n",
              "      progress.max = size;\n",
              "      div.appendChild(progress);\n",
              "      document.body.appendChild(div);\n",
              "\n",
              "      const buffers = [];\n",
              "      let downloaded = 0;\n",
              "\n",
              "      const channel = await google.colab.kernel.comms.open(id);\n",
              "      // Send a message to notify the kernel that we're ready.\n",
              "      channel.send({})\n",
              "\n",
              "      for await (const message of channel.messages) {\n",
              "        // Send a message to notify the kernel that we're ready.\n",
              "        channel.send({})\n",
              "        if (message.buffers) {\n",
              "          for (const buffer of message.buffers) {\n",
              "            buffers.push(buffer);\n",
              "            downloaded += buffer.byteLength;\n",
              "            progress.value = downloaded;\n",
              "          }\n",
              "        }\n",
              "      }\n",
              "      const blob = new Blob(buffers, {type: 'application/binary'});\n",
              "      const a = document.createElement('a');\n",
              "      a.href = window.URL.createObjectURL(blob);\n",
              "      a.download = filename;\n",
              "      div.appendChild(a);\n",
              "      a.click();\n",
              "      div.remove();\n",
              "    }\n",
              "  "
            ],
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "download(\"download_c5d7d86b-a78f-4a7c-8f7c-bae8326d85ec\", \"C_FM_1.hkl\", 6224)"
            ],
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "\n",
              "    async function download(id, filename, size) {\n",
              "      if (!google.colab.kernel.accessAllowed) {\n",
              "        return;\n",
              "      }\n",
              "      const div = document.createElement('div');\n",
              "      const label = document.createElement('label');\n",
              "      label.textContent = `Downloading \"${filename}\": `;\n",
              "      div.appendChild(label);\n",
              "      const progress = document.createElement('progress');\n",
              "      progress.max = size;\n",
              "      div.appendChild(progress);\n",
              "      document.body.appendChild(div);\n",
              "\n",
              "      const buffers = [];\n",
              "      let downloaded = 0;\n",
              "\n",
              "      const channel = await google.colab.kernel.comms.open(id);\n",
              "      // Send a message to notify the kernel that we're ready.\n",
              "      channel.send({})\n",
              "\n",
              "      for await (const message of channel.messages) {\n",
              "        // Send a message to notify the kernel that we're ready.\n",
              "        channel.send({})\n",
              "        if (message.buffers) {\n",
              "          for (const buffer of message.buffers) {\n",
              "            buffers.push(buffer);\n",
              "            downloaded += buffer.byteLength;\n",
              "            progress.value = downloaded;\n",
              "          }\n",
              "        }\n",
              "      }\n",
              "      const blob = new Blob(buffers, {type: 'application/binary'});\n",
              "      const a = document.createElement('a');\n",
              "      a.href = window.URL.createObjectURL(blob);\n",
              "      a.download = filename;\n",
              "      div.appendChild(a);\n",
              "      a.click();\n",
              "      div.remove();\n",
              "    }\n",
              "  "
            ],
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "download(\"download_a22c80ad-0c2b-4efd-a00f-178eef1a20e0\", \"I_FM_1.hkl\", 6432)"
            ],
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Xy8iq8iS02dv"
      },
      "source": [
        "# Unlock Phase-2"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "G9p6_IYDSlyP",
        "outputId": "1708080c-b86d-4cdd-d50e-77300c55e86b"
      },
      "source": [
        "%reset -f\n",
        "!pip install faiss-gpu\n",
        "!pip install celluloid\n",
        "\n",
        "import faiss\n",
        "import numpy as np\n",
        "from scipy.stats import norm\n",
        "import time\n",
        "import torch\n",
        "from google.colab import files\n",
        "import random\n",
        "from sklearn.neighbors import NearestNeighbors\n",
        "device = 'cuda'\n",
        "res = faiss.StandardGpuResources()\n",
        "select = 0\n",
        "MonteCarlo = 2\n",
        "# Household positions of each agent\n",
        "Residence = torch.tensor(np.loadtxt('mumbai_agent_location.dat'),dtype=torch.double,device='cuda')\n",
        "# Age of the individual\n",
        "Age = torch.tensor(np.loadtxt('mumbai_Age.dat'),dtype=torch.long,device='cuda')\n",
        "# Household size\n",
        "H_Size = torch.tensor(np.loadtxt('mumbai_Age_House.dat'),dtype=torch.long,device='cuda')\n",
        "# 1-member households\n",
        "H_Size[H_Size==0] = 1\n",
        "# ID of the households of each agent\n",
        "H_Number = torch.tensor(np.loadtxt('mumbai_House_number.dat'),dtype=torch.long,device='cuda')\n",
        "# Location of each household\n",
        "H_location = torch.tensor(np.loadtxt('mumbai_household_location.dat'),dtype=torch.long,device='cuda')\n",
        "# Total Population\n",
        "Population = 206340\n",
        "# Number of Children (0), Adults (1), Seniors (2) and Infections (3)\n",
        "Demography = torch.tensor([round(0.2206*Population),round(0.7212*Population),0,round(0.0004345*Population),685],dtype=torch.long,device='cuda')\n",
        "Demography[2] = Population-torch.sum(Demography[:2]) \n",
        "Demography[3] = 800\n",
        "# Layout Size: XMinimum (0), XMaximum (1), YMinimum (2), YMaximum (3), Grid Resolution (4)\n",
        "Layout = torch.tensor([0,5000,0,2000,1000],dtype=torch.double,device='cuda')\n",
        "# Minimum (0), Maximum (1), Inner (2), Innermost (3) Allowed Speeds\n",
        "VelocityParameters = torch.tensor([-500,500,200,10,1],dtype=torch.long,device='cuda')\n",
        "# Collision Radius (0), Infection Radius (1), OuterCircle (2), InnerCircle (3), Reached Goal (4)\n",
        "Radii = torch.tensor([5,2,570,120,5],dtype=torch.long,device='cuda') # Infection Radius(20) > Collision Radius(5/50) --> no social distancing\n",
        "# Total time (days), Resolution(minutes)\n",
        "Time = torch.tensor([31,15],dtype=torch.long, device='cuda') \n",
        "# TimeSteps in a day (1 step = 15mins) and total\n",
        "TimeStepsDay = torch.div(24*60,Time[1].item())\n",
        "# Total time steps in 31 days\n",
        "TimeStepsTotal = TimeStepsDay*Time[0].long()\n",
        "# Re-assign x and y initial positions\n",
        "# Assign positions within unit radius\n",
        "adjustment = 0.5\n",
        "Residence = H_location[H_Number] + torch.DoubleTensor(Population,2).uniform_(-adjustment,adjustment).cuda()\n",
        "# Check if the newly assigned x co-ordinate is within the boundary\n",
        "Residence[Residence[:,0]>Layout[1],0] = (torch.DoubleTensor(Population).uniform_(Layout[1]-adjustment,Layout[1]).cuda())[Residence[:,0]>Layout[1]]\n",
        "Residence[Residence[:,0]<Layout[0],0] = (torch.DoubleTensor(Population).uniform_(Layout[0],Layout[0]+adjustment).cuda())[Residence[:,0]<Layout[0]]\n",
        "# Check if the newly assigned y co-ordinate is within the boundary\n",
        "Residence[Residence[:,1]>Layout[3],1] = (torch.DoubleTensor(Population).uniform_(Layout[3]-adjustment,Layout[3]).cuda())[Residence[:,1]>Layout[3]]\n",
        "Residence[Residence[:,1]<Layout[2],1] = (torch.DoubleTensor(Population).uniform_(Layout[2],Layout[2]+adjustment).cuda())[Residence[:,1]<Layout[2]]\n",
        "Activities = np.zeros((20,96),dtype=np.float32)\n",
        "# Children\n",
        "# Errands \n",
        "Activities[0,24:36]=0.001*norm.pdf(np.arange(6,9,step=0.25), 7.5, 1) # 6-9\n",
        "Activities[0,36:56]=0.002*norm.pdf(np.arange(9,14,step=0.25), 11.5, 1) # 9-14\n",
        "Activities[0,56:72]=0.003*norm.pdf(np.arange(14,18,step=0.25), 17, 1) # 14-21\n",
        "# Adults\n",
        "# Errands \n",
        "Activities[8,24:36]=0.005*norm.pdf(np.arange(6,9,step=0.25), 7.5, 1) # 6-9\n",
        "Activities[8,36:56]=0.01*norm.pdf(np.arange(9,14,step=0.25), 11.5, 1) # 9-14\n",
        "Activities[8,56:72]=0.02*norm.pdf(np.arange(14,18,step=0.25), 17, 1) # 14-21\n",
        "# Nomadic\n",
        "Activities[12,24:48] = 0.05*norm.pdf(np.arange(6,12,step=0.25), 10, 1) # 6-12\n",
        "Activities[12,48:72] = 0.1*norm.pdf(np.arange(12,18,step=0.25), 16, 1) # 12-21\n",
        "# Nomads go home (18) # Weekday\n",
        "Activities[18,0:24] = np.ones(24)\n",
        "Activities[18,84:96] = np.ones(12)\n",
        "# Nomads go home (19) # Weekend\n",
        "Activities[19,0:96] = np.ones(96)\n",
        "# Children\n",
        "# Market\n",
        "Activities[1,24:48]=0.001*norm.pdf(np.arange(6,12,step=0.25), 10, 1) # 6-12\n",
        "Activities[1,48:72]=0.002*norm.pdf(np.arange(12,18,step=0.25), 16, 1) # 12-21\n",
        "# Adults\n",
        "# Market\n",
        "Activities[9,24:48]=0.005*norm.pdf(np.arange(6,12,step=0.25), 10, 1) # 6-12\n",
        "Activities[9,48:72]=0.01*norm.pdf(np.arange(12,18,step=0.25), 17, 1) # 12-21\n",
        "# Office \n",
        "Activities[10,24:48]= 2*norm.pdf(np.arange(6,12,step=0.25), 9, 1) # 6-12\n",
        "Activities[11,48:72] = 2*norm.pdf(np.arange(12,18,step=0.25), 16, 1) # 12-21\n",
        "\n",
        "\n",
        "# Weekday Activities\n",
        "Limits = torch.tensor([0,0,0,0,0,0,1,100,1,100,0,0,\n",
        "                       0,0,0,0,0,0,0,0,0,0,0,0,\n",
        "                       0,0,0,0,2,100,0,100,0,100,0,0,\n",
        "                       0,100,0,0,0,100,0,100,0,100,0,0,\n",
        "                       15,100,0,0,0,100,0,100,0,100,0,100,\n",
        "                       0,0,0,0,0,0,0,0,0,100,0,100],dtype=torch.long,device='cuda')\n",
        "# Weekend Activities                    \n",
        "Limits_Weekend = torch.tensor([0,0,0,0,0,0,0,0,0,0,0,0,\n",
        "                               0,0,0,0,0,0,0,0,0,0,0,0,\n",
        "                               0,0,0,0,0,0,0,0,0,0,0,0,\n",
        "                               0,0,0,0,0,0,0,0,0,0,0,0,\n",
        "                               0,0,0,0,0,0,0,0,0,0,0,0,\n",
        "                               0,0,0,0,0,0,0,0,0,0,0,0],dtype=torch.long,device='cuda')\n",
        "Location_one = torch.tensor([12,13,14,15,16,17],dtype=torch.long,device='cuda') # Location of i\n",
        "Location_two = torch.tensor([[12,0,0,15,16,0],\n",
        "                             [0,0,0,0,0,0],\n",
        "                             [0,0,14,15,16,0],\n",
        "                             [12,0,14,15,16,0],\n",
        "                            [12,0,14,15,16,17],\n",
        "                             [0,0,0,0,16,17]],dtype=torch.long,device='cuda') # Corresponding locations of j  \n",
        "\n",
        "def goal_assign(x,y,minbias,maxbias,xhalfmark,yhalfmark):\n",
        "  '''\n",
        "  Generate Random Goals\n",
        "  '''\n",
        "  # ordinate\n",
        "  l = torch.zeros((Population),dtype=torch.double,device='cuda')\n",
        "  # abscissa\n",
        "  m = torch.zeros((Population),dtype=torch.double,device='cuda')\n",
        "  temp = x -  torch.FloatTensor(Population).uniform_(minbias,maxbias).cuda()\n",
        "  l[x>xhalfmark] = temp[x>xhalfmark] \n",
        "  temp = y -  torch.FloatTensor(Population).uniform_(minbias,maxbias).cuda()\n",
        "  m[y>yhalfmark] = temp[y>yhalfmark] \n",
        "  temp = x +  torch.FloatTensor(Population).uniform_(minbias,maxbias).cuda()\n",
        "  l[x<xhalfmark] = temp[x<xhalfmark] \n",
        "  temp = y +  torch.FloatTensor(Population).uniform_(minbias,maxbias).cuda()\n",
        "  m[y<yhalfmark] = temp[y<yhalfmark] \n",
        "  return l,m   \n",
        "\n",
        "def generateMatrix(agents,matrixone,limits,Population, H_Number, index,con_one, con_two,Time_Daily,matrixthree,distance,radius):\n",
        "  '''\n",
        "  Generate a conditional contact matrix\n",
        "  '''\n",
        "  # Store end result\n",
        "  matrixtwo = torch.zeros((2,Population,20),dtype=torch.double,device='cuda')\n",
        "  count=0\n",
        "  for i in range(6):\n",
        "    vec_con1 = torch.zeros((Population),device='cuda',dtype=torch.bool)\n",
        "    vec_con1 += (agents==con_one[i])\n",
        "    Time_Daily[vec_con1==True,i]+=1\n",
        "    for j in range(6):\n",
        "      temp = torch.zeros((Population,20),device='cuda',dtype=torch.int32)\n",
        "      vec_con2 = torch.zeros((Population,20),device='cuda',dtype=torch.bool)      \n",
        "      if (con_two[i][j]!=0):\n",
        "        if (count==0):          \n",
        "          vec_con2 = torch.logical_and(H_Number[index]==H_Number[:,None],(agents[index]==con_two[i][j]))\n",
        "        else:\n",
        "          vec_con2 = (agents[index]==con_two[i][j])\n",
        "        temp = vec_con1[:,None]*vec_con2\n",
        "        matrixone[:,count] += torch.sum(temp,1)\n",
        "        #matrixone[:,count] += torch.sum(torch.logical_and(temp,distance<radius),1)\n",
        "        if (count!=0):\n",
        "          matrixone[matrixone[:,count]>limits[:,count],count] = limits[matrixone[:,count]>limits[:,count],count]\n",
        "        #tempone = temp*(1-(torch.cumsum(torch.logical_and(temp,distance<radius),axis=1)>matrixone[:,count][:,None]).long())\n",
        "        tempone = temp*(1-(torch.cumsum(temp,axis=1)>matrixone[:,count][:,None]).long())\n",
        "        matrixtwo[0] += tempone\n",
        "        matrixtwo[1] = temp\n",
        "        if (torch.where(vec_con1==True)[0].size(0)!=0):\n",
        "          matrixthree[i,j] += torch.where(tempone==1)[0].size(0)/torch.where(vec_con1==True)[0].size(0)\n",
        "      count+=1\n",
        "  return matrixtwo[0],matrixtwo[1], matrixone, matrixthree, Time_Daily\n",
        "\n",
        "def nearest_destination(Population,centroids,Residence): \n",
        "  '''\n",
        "  Find the nearest destination: Office/School/Market\n",
        "  '''\n",
        "  distance = torch.zeros((Population,len(centroids)),device='cuda',dtype=torch.double)\n",
        "  for i in range(len(centroids)):\n",
        "    # Check the distance from the centroid of the zone to each household location\n",
        "    distance[:,i] = torch.linalg.norm(Residence-Centroids[i],axis=1)\n",
        "  return torch.argmin(distance,axis=1)\n",
        "# Schools\n",
        "# Co-ordinates of different schools\n",
        "School = torch.tensor([[1000,1500,300,500],\n",
        "                       [2000,2500,300,500],\n",
        "                       [3000,3600,500,1000],\n",
        "                       [3500,4000,1300,1500],\n",
        "                       [600,1000,1100,1500],\n",
        "                       [2400,2800,1600,1800]],dtype=torch.long,device='cuda')\n",
        "# Generate Destinations in different school zones\n",
        "School_Positions = torch.zeros((Population,12),dtype=torch.double,device='cuda')\n",
        "# x-co-ordinate\n",
        "School_Positions[:,0] = torch.DoubleTensor(Population).uniform_(School[0,0],School[0,1])\n",
        "School_Positions[:,2] = torch.DoubleTensor(Population).uniform_(School[1,0],School[1,1])\n",
        "School_Positions[:,4] = torch.DoubleTensor(Population).uniform_(School[2,0],School[2,1])\n",
        "School_Positions[:,6] = torch.DoubleTensor(Population).uniform_(School[3,0],School[3,1])\n",
        "School_Positions[:,8] = torch.DoubleTensor(Population).uniform_(School[4,0],School[4,1])\n",
        "School_Positions[:,10] = torch.DoubleTensor(Population).uniform_(School[5,0],School[5,1])\n",
        "# y-co-ordinate\n",
        "School_Positions[:,1] = torch.DoubleTensor(Population).uniform_(School[0,2],School[0,3])\n",
        "School_Positions[:,3] = torch.DoubleTensor(Population).uniform_(School[1,2],School[1,3])\n",
        "School_Positions[:,5] = torch.DoubleTensor(Population).uniform_(School[2,2],School[2,3])\n",
        "School_Positions[:,7] = torch.DoubleTensor(Population).uniform_(School[3,2],School[3,3])\n",
        "School_Positions[:,9] = torch.DoubleTensor(Population).uniform_(School[4,2],School[4,3])\n",
        "School_Positions[:,11] = torch.DoubleTensor(Population).uniform_(School[5,2],School[5,3])\n",
        "\n",
        "# Offices\n",
        "# Co-ordinates of different offices\n",
        "Office = torch.tensor([[1000,1500,0,300],\n",
        "                      [2000,2500,0,300],\n",
        "                      [3500,4000,1000,1300],\n",
        "                      [2000,3000,500,1000]],dtype=torch.long,device='cuda')\n",
        "# Generate destinations in office zones\n",
        "Office_Positions = torch.zeros((Population,8),dtype=torch.double,device='cuda')\n",
        "# x co-ordinate\n",
        "Office_Positions[:,0] = torch.DoubleTensor(Population).uniform_(Office[0,0],Office[0,1])\n",
        "Office_Positions[:,2] = torch.DoubleTensor(Population).uniform_(Office[1,0],Office[1,1])\n",
        "Office_Positions[:,4] = torch.DoubleTensor(Population).uniform_(Office[2,0],Office[2,1])\n",
        "Office_Positions[:,6] = torch.DoubleTensor(Population).uniform_(Office[3,0],Office[3,1])\n",
        "# y co-ordinate\n",
        "Office_Positions[:,1] = torch.DoubleTensor(Population).uniform_(Office[0,2],Office[0,3])\n",
        "Office_Positions[:,3] = torch.DoubleTensor(Population).uniform_(Office[1,2],Office[1,3])\n",
        "Office_Positions[:,5] = torch.DoubleTensor(Population).uniform_(Office[2,2],Office[2,3])\n",
        "Office_Positions[:,7] = torch.DoubleTensor(Population).uniform_(Office[3,2],Office[3,3])\n",
        "\n",
        "# Markets\n",
        "# Co-ordinates of different markets\n",
        "Market = torch.tensor([[2600,3000,1500,2000],\n",
        "                       [3600,4200,100,400],\n",
        "                       [4600,5000,700,900],\n",
        "                       [2000,2400,1000,1200],\n",
        "                       [0,400,1000,1300],\n",
        "                       [3600,4000,1500,1700],\n",
        "                       [1000,2000,500,1000],\n",
        "                       [300,900,1600,1800]],dtype=torch.long,device='cuda')\n",
        "# Generate destinations in market zones\n",
        "Market_Positions = torch.zeros((Population,16),dtype=torch.double,device='cuda')\n",
        "# x co-ordinate\n",
        "Market_Positions[:,0] = torch.DoubleTensor(Population).uniform_(Market[0,0],Market[0,1])\n",
        "Market_Positions[:,2] = torch.DoubleTensor(Population).uniform_(Market[1,0],Market[1,1])\n",
        "Market_Positions[:,4] = torch.DoubleTensor(Population).uniform_(Market[2,0],Market[2,1])\n",
        "Market_Positions[:,6] = torch.DoubleTensor(Population).uniform_(Market[3,0],Market[3,1])\n",
        "Market_Positions[:,8] = torch.DoubleTensor(Population).uniform_(Market[4,0],Market[4,1])\n",
        "Market_Positions[:,10] = torch.DoubleTensor(Population).uniform_(Market[5,0],Market[5,1])\n",
        "Market_Positions[:,12] = torch.DoubleTensor(Population).uniform_(Market[6,0],Market[6,1])\n",
        "Market_Positions[:,14] = torch.DoubleTensor(Population).uniform_(Market[7,0],Market[7,1])\n",
        "# y co-ordinate\n",
        "Market_Positions[:,1] = torch.DoubleTensor(Population).uniform_(Market[0,2],Market[0,3])\n",
        "Market_Positions[:,3] = torch.DoubleTensor(Population).uniform_(Market[1,2],Market[1,3])\n",
        "Market_Positions[:,5] = torch.DoubleTensor(Population).uniform_(Market[2,2],Market[2,3])\n",
        "Market_Positions[:,7] = torch.DoubleTensor(Population).uniform_(Market[3,2],Market[3,3])\n",
        "Market_Positions[:,9] = torch.DoubleTensor(Population).uniform_(Market[4,2],Market[4,3])\n",
        "Market_Positions[:,11] = torch.DoubleTensor(Population).uniform_(Market[5,2],Market[5,3])\n",
        "Market_Positions[:,13] = torch.DoubleTensor(Population).uniform_(Market[6,2],Market[6,3])\n",
        "Market_Positions[:,15] = torch.DoubleTensor(Population).uniform_(Market[7,2],Market[7,3])\n",
        "\n",
        "# Centroids of each zone\n",
        "Centroids = torch.zeros((18,2),dtype=torch.double,device='cuda')\n",
        "# Centroids of school zones\n",
        "Centroids[:6,0] = School[:,0] + 0.5*School[:,1]\n",
        "Centroids[:6,1] = School[:,2] + 0.5*School[:,3]\n",
        "# Centroids of office zones\n",
        "Centroids[6:10,0] = Office[:,0] + 0.5*Office[:,1]\n",
        "Centroids[6:10,1] = Office[:,2] + 0.5*Office[:,3]\n",
        "# Centroids of market zones\n",
        "Centroids[10:18,0] = Market[:,0] + 0.5*Market[:,1]\n",
        "Centroids[10:18,1] = Market[:,2] + 0.5*Market[:,3]\n",
        "\n",
        "# Find nearest destination\n",
        "Nearest = torch.zeros((Population,3),device='cuda',dtype=torch.long)\n",
        "# Nearest school\n",
        "Nearest[:,0] = nearest_destination(Population,Centroids[:6],Residence)\n",
        "# Nearest office\n",
        "Nearest[:,1] = nearest_destination(Population,Centroids[6:10],Residence)\n",
        "# Nearest market\n",
        "Nearest[:,2] = nearest_destination(Population,Centroids[10:18],Residence) \n",
        "# ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n",
        "# Disease Transmission Parameters\n",
        "# -------------------------------------\n",
        "\n",
        "# ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n",
        "# Disease Transmission Parameters\n",
        "# -------------------------------------\n",
        "\n",
        "# States: SIT/SIN/AIT/AIN/SUT/SUN\n",
        "States = torch.tensor(([0,0,0,0,0,1,2,3],\n",
        "                       [0,0,0,1,0,0,0,0],\n",
        "                       [6,7,6,7,0,0,0,0],\n",
        "                       [9,9,9,9,0,0,0,0],\n",
        "                       [0,0,0,0,0,0,0,0],\n",
        "                       [0,1,2,3,4,5,6,7]),dtype=torch.int,device='cuda')\n",
        "# Disease Transition Probabilities\n",
        "Transition = torch.zeros((3,6,8),dtype=torch.double,device='cuda')\n",
        "undercount = 0\n",
        "# Children \n",
        "Transition[0] = torch.transpose(torch.tensor(([0,0,0.12495,0.00005,0,0.875],\n",
        "                                              [0,0,0.09994,0.00006,undercount,0.9-undercount],\n",
        "                                              [0,0.00998,0.19,0.00002,0,0.8],\n",
        "                                              [0,0.02996,0.12,0.00004,undercount,0.85-undercount],\n",
        "                                              [0.2537,0,0,0,0,0.7463],\n",
        "                                              [0.2537,0,0,0,0,0.7463],\n",
        "                                              [0.2537,0,0,0,0,0.7463],\n",
        "                                              [0.2537,0,0,0,0,0.7463]),dtype=torch.double,device='cuda'),0,1) \n",
        "# Adults\n",
        "Transition[1] = torch.transpose(torch.tensor(([0,0,0.124,0.001,0,0.875],\n",
        "                                              [0,0,0.06,0.04,undercount,0.9-undercount],\n",
        "                                              [0,0.03993,0.16,0.00007,0,0.8],\n",
        "                                              [0,0.0493,0.1,0.0007,undercount,0.85-undercount],\n",
        "                                              [0.5,0,0,0,0,0.5],\n",
        "                                              [0.5,0,0,0,0,0.5],\n",
        "                                              [0.5,0,0,0,0,0.5],\n",
        "                                              [0.5,0,0,0,0,0.5]),dtype=torch.double,device='cuda'),0,1) \n",
        "# Seniors\n",
        "Transition[2] = torch.transpose(torch.tensor(([0,0,0.065,0.06,0,0.875],\n",
        "                                              [0,0,0.02,0.08,undercount,0.9-undercount],\n",
        "                                              [0,0.08,0.1,0.02,0,0.8],\n",
        "                                              [0,0.1,0.01,0.04,undercount,0.85-undercount],\n",
        "                                              [0.5951,0,0,0,0,0.4049],\n",
        "                                              [0.5951,0,0,0,0,0.4049],\n",
        "                                              [0.5951,0,0,0,0,0.4049],\n",
        "                                              [0.5951,0,0,0,0,0.4049]),dtype=torch.double,device='cuda'),0,1)\n",
        "# Storage Parameters\n",
        "# Outermost Loop (Scenario/Final Parameters)\n",
        "# Disease Transmission: S/IT/INT/R/D/Total (7+7, Mean and Variance states) \n",
        "FinalTimeSeries = torch.zeros((65,(TimeStepsTotal.long().item())+1), dtype=torch.float,device='cuda')\n",
        "# Time to infection, death, recovery (Mean - 0,1,2, Variance - 3,4,5)\n",
        "FinalConstants = torch.zeros((20), dtype=torch.float,device='cuda')\n",
        "InteractionsFinal = torch.zeros((2,6,6),dtype=torch.float,device='cuda')\n",
        "# Outermost loop\n",
        "# Second Loop (MonteCarlo Run Parameters)\n",
        "# ResultsTimeSeries (Time Series for MC runs)(0-6)\n",
        "ResultsTMC = torch.zeros((MonteCarlo,29,(TimeStepsTotal.long().item())+1), dtype=torch.float,device='cuda')\n",
        "# Results Constants: Time to Infection: Overall(0), Recovery(1), Death(2)\n",
        "ResultsCMC = torch.zeros((MonteCarlo,10,1), dtype=torch.float,device='cuda')\n",
        "InteractionsMC = torch.zeros((MonteCarlo,2,6,6), dtype=torch.float,device='cuda')\n",
        "for mc in range(MonteCarlo):\n",
        "  # Create N agents;\n",
        "  # Age = 0,     \n",
        "  # x_position = 1, \n",
        "  # y_position = 2, \n",
        "  # x_velocity = 3, \n",
        "  # y_velocity = 4, \n",
        "  # occupation = 5,\n",
        "  # current_task = 7, \n",
        "  # status_of_task = 8, \n",
        "  # goal_x = 9, \n",
        "  # goal_y = 10,\n",
        "  # infection_status = 14, \n",
        "  # infection_state = 15, \n",
        "  # time_of_infectiousness = 16\n",
        "  Agents = torch.zeros((Population,20),dtype=torch.double,device='cuda')\n",
        "  # Assign Age \n",
        "  Agents[:,0] = Age\n",
        "  # Indices of Children\n",
        "  Children = torch.where(Agents[:,0]==0)[0]\n",
        "  # Indices of Adults\n",
        "  Adults = torch.where(Agents[:,0]==1)[0]\n",
        "  # Indices of Seniors\n",
        "  Seniors = torch.where(Agents[:,0]==2)[0]\n",
        "  # Assign x and y Positions\n",
        "  # PositionX (1)\n",
        "  Agents[:,1] = Residence[:,0].detach().clone()\n",
        "  # PositionY (2)\n",
        "  Agents[:,2] = Residence[:,1].detach().clone()\n",
        "  # Assign x and y Velocity Components\n",
        "  # VelocityX (3)\n",
        "  Agents[:,3] = torch.DoubleTensor(Population).uniform_(VelocityParameters[0],VelocityParameters[1])\n",
        "  # VelocityY (4)\n",
        "  Agents[:,4] = torch.DoubleTensor(Population).uniform_(VelocityParameters[0],VelocityParameters[1])\n",
        "  # Social Distance:  \n",
        "  SD = Radii[0]*torch.ones((Population),dtype=torch.long,device='cuda')\n",
        "  # Fraction not following (0: Everyone, 1: No one)\n",
        "  SD[torch.tensor([True,False],device='cuda')[torch.multinomial(torch.tensor([0.9,0.1]),Population,replacement=True)]] = 0\n",
        "  # Occupation (5): \n",
        "  # School-going Children (1)\n",
        "  Agents[Children[:int(0.9*Demography[0])],5] = 1\n",
        "  # School-going Adults (1)\n",
        "  Agents[Adults[:int(0.05*Demography[1])],5] = 1\n",
        "  # School-going Seniors (1)\n",
        "  Agents[Seniors[:int(0.02*Demography[2])],5] = 1\n",
        "  # Stay-at-home adults (2)\n",
        "  Agents[Adults[int(0.05*Demography[1]):int(0.25*Demography[1])],5] = 2\n",
        "  # Office Going Adults (3)\n",
        "  Agents[Adults[int(0.25*Demography[1]):int(0.85*Demography[1])],5] = 3\n",
        "  # Nomadic Adults (4)\n",
        "  Agents[Adults[int(0.85*Demography[1]):Demography[1]],5] = 4\n",
        "  # Stay-at-home Seniors (5)\n",
        "  Agents[Seniors[int(0.02*Demography[2]):int(0.62*Demography[2])],5] = 5 \n",
        "  # Office Going Seniors (6)\n",
        "  Agents[Seniors[int(0.62*Demography[2]):int(0.95*Demography[2])],5] = 6\n",
        "  # Reduce office-goers by fraction\n",
        "  # Nomadic Seniors (7)\n",
        "  Agents[Seniors[int(0.95*Demography[2]):Demography[2]],5] = 7\n",
        "  # Assign Infection Status: Infected and Tested (1), Uninfect(0), Infected and Not Tested (3)\n",
        "  # 0.5% infections among children, 49% infections among adults, 50.5% among seniors\n",
        "  Agents[Children[:int(0.005*Demography[3])],14] = 3 # Infected \n",
        "  Agents[Adults[int(0.005*Demography[3]):int(0.49*Demography[3])],14] = 3 # Infected \n",
        "  Agents[Seniors[int(0.49*Demography[3]):Demography[3]],14] = 3 # Infected \n",
        "  Agents[torch.where(Agents[:,14]==3)[0][:int(Demography[4])],14] = 1 # Confirmed Infections\n",
        "  # Assign Infection State: SIT=0, SIN=1, AIT=2, AIN=3, SUT=4, SUN=5, AUT=6, AUN=7\n",
        "  # Infected\n",
        "  # 80% of the infections are asymptomatic\n",
        "  # 20% of the infections are symptomatic\n",
        "  Agents[torch.where(Agents[:,14]==3)[0][:int(0.8*Demography[3])],15] = 3 # AIN\n",
        "  Agents[torch.where(Agents[:,14]==3)[0][int(0.8*Demography[3]):Demography[3]],15] = 1 # SIN\n",
        "  # Uninfected\n",
        "  # 85% of the uninfected are asymptomatic\n",
        "  # 15% of the uninfected are symptomatic\n",
        "  Agents[torch.where(Agents[:,14]==0)[0][:int(0.85*(Population-Demography[3]))],15] = 7 # AUN\n",
        "  Agents[torch.where(Agents[:,14]==0)[0][int(0.85*(Population-Demography[3])):Population-Demography[3]],15] = 5 # SUN\n",
        "  Agents[torch.where(Agents[:,14]==1)[0],15] = 2\n",
        "  # --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n",
        "  # Placeholder Variables\n",
        "  # -------------------------------------\n",
        "\n",
        "  # Vectors \n",
        "  Vectors = torch.zeros((12,Population),device='cuda',dtype=torch.double)\n",
        "  # Distance Matrix(0), Index Matrix(1), Condition Matrix(2)\n",
        "  Matrixs = torch.zeros((5,Population,20),dtype=torch.double,device='cuda')\n",
        "  # Initial Position \n",
        "  CurrentPosition = torch.dstack((Agents[:,1],Agents[:,2]))\n",
        "  # Initial Velocity\n",
        "  Velocity = torch.dstack((Agents[:,3],Agents[:,4]))\n",
        "  # All agents are initially at home (12)\n",
        "  Agents[:,7] = 12\n",
        "  # Status of the task is complete and initially at home (12)\n",
        "  Agents[:,8] = Agents[:,7].detach().clone()\n",
        "  # Ordinate of goal position\n",
        "  Agents[:,9] = Agents[:,1].detach().clone()\n",
        "  # Abscissa of goal position\n",
        "  Agents[:,10] = Agents[:,2].detach().clone()\n",
        "  # Initial Goals are to remain at home\n",
        "  Destination = torch.dstack((Agents[:,1],Agents[:,2]))\n",
        "  # Days \n",
        "  Days = torch.tensor([0],device='cuda')       \n",
        "  # Distance\n",
        "  Distance = torch.zeros((Population),dtype=torch.double,device='cuda')\n",
        "  # Infection Status over Time\n",
        "  # Overall\n",
        "  Results = torch.tensor([torch.where(Agents[:,14]==0)[0].size()[0], #0:Uninfected\n",
        "                          torch.where(Agents[:,14]==1)[0].size()[0], #1:Infected and Tested\n",
        "                          torch.where(Agents[:,14]==2)[0].size()[0], #2:Unreported Recoveries\n",
        "                          torch.where(Agents[:,14]==3)[0].size()[0], #3:Infected and Not Tested\n",
        "                          torch.where(Agents[:,14]==4)[0].size()[0], #4:Unreported Deaths\n",
        "                          torch.where(Agents[:,14]==5)[0].size()[0], #5:Reported Recoveries\n",
        "                          torch.where(Agents[:,14]==6)[0].size()[0]],dtype=torch.long,device='cuda') #6:Reported Deaths  \n",
        "  # Children\n",
        "  ResultsC = torch.tensor([torch.where(Agents[Children,14]==0)[0].size()[0],\n",
        "                          torch.where(Agents[Children,14]==1)[0].size()[0],\n",
        "                          torch.where(Agents[Children,14]==2)[0].size()[0],\n",
        "                          torch.where(Agents[Children,14]==3)[0].size()[0],\n",
        "                          torch.where(Agents[Children,14]==4)[0].size()[0],\n",
        "                          torch.where(Agents[Children,14]==5)[0].size()[0],\n",
        "                          torch.where(Agents[Children,14]==6)[0].size()[0]],dtype=torch.long,device='cuda')   \n",
        "  # Adults  \n",
        "  ResultsA = torch.tensor([torch.where(Agents[Adults,14]==0)[0].size()[0],\n",
        "                          torch.where(Agents[Adults,14]==1)[0].size()[0],\n",
        "                          torch.where(Agents[Adults,14]==2)[0].size()[0],\n",
        "                          torch.where(Agents[Adults,14]==3)[0].size()[0],\n",
        "                          torch.where(Agents[Adults,14]==4)[0].size()[0],\n",
        "                          torch.where(Agents[Adults,14]==5)[0].size()[0],\n",
        "                          torch.where(Agents[Adults,14]==6)[0].size()[0]],dtype=torch.long,device='cuda')   \n",
        "  # Seniors\n",
        "  ResultsS = torch.tensor([torch.where(Agents[Seniors,14]==0)[0].size()[0],\n",
        "                          torch.where(Agents[Seniors,14]==1)[0].size()[0],\n",
        "                          torch.where(Agents[Seniors,14]==2)[0].size()[0],\n",
        "                          torch.where(Agents[Seniors,14]==3)[0].size()[0],\n",
        "                          torch.where(Agents[Seniors,14]==4)[0].size()[0],\n",
        "                          torch.where(Agents[Seniors,14]==5)[0].size()[0],\n",
        "                          torch.where(Agents[Seniors,14]==6)[0].size()[0]],dtype=torch.long,device='cuda')   \n",
        "\n",
        "  # Interactions: Weekday=0, Weekend=1, Contact_Matrix=2\n",
        "  Interactions = torch.zeros((3,6,6),dtype=torch.double,device='cuda')\n",
        "  # Weekday = 0, Weekend = 1\n",
        "  Time_avg = torch.zeros((2,6),dtype=torch.double,device='cuda')\n",
        "  Time_Daily = torch.zeros((Population,6),dtype=torch.long,device='cuda')\n",
        "  # Number of contacts on a day to day basis: Daily_Contact=0, Contact_Limit=1\n",
        "  Contacts = torch.zeros((2,Population,36),dtype=torch.long,device='cuda')\n",
        "\n",
        "  for t in range(TimeStepsTotal.long().item()):\n",
        "\n",
        "    Days = torch.floor(torch.div(t,TimeStepsDay).cuda())\n",
        "    # Time of the day\n",
        "    time_of_day = t%96 \n",
        "    # Day of the week\n",
        "    day_of_week = Days%7 \n",
        "\n",
        "    # ------------------------------------------------\n",
        "    # Store parameters from the previous run\n",
        "    # ------------------------------------------------\n",
        "\n",
        "    # Infection Status from previous iteration\n",
        "    Vectors[4] = Agents[:,14].detach().clone()\n",
        "    # Infection State from previous iteration\n",
        "    Vectors[5] = Agents[:,15].detach().clone()\n",
        "    # Rtime\n",
        "    Vectors[6] = Vectors[1].detach().clone()\n",
        "\n",
        "    # ---------------------------------------------\n",
        "    # Find Limits for the day\n",
        "    # ---------------------------------------------\n",
        "\n",
        "    if time_of_day==0:\n",
        "      # Assign contact limits for the weekend\n",
        "      if torch.logical_or(day_of_week==5,day_of_week==6):\n",
        "        for i in range(1,36):\n",
        "          Contacts[1][:,i] = Limits_Weekend[2*i+1]\n",
        "      else:\n",
        "        # Assign contact limits for the weekday\n",
        "        for i in range(1,36):\n",
        "          Contacts[1][:,i] = Limits[2*i+1]  \n",
        "\n",
        "    #Agents[:,15]: SIT=0, SIN=1, AIT=2, AIN=3, SUT=4, SUN=5, AUT=6, AUN=7\n",
        "    #Unconfirmed to Confirmed (30 times underreported)\n",
        "    Vectors[0] = torch.rand(Population,device='cuda').double()< 0.000010627#0.00000726 #0.0000450422\n",
        "    Vectors[7] = torch.logical_and(Agents[:,15]==1,Vectors[0]==True)\n",
        "    Agents[Vectors[7].bool(),15] = 0 # SIN --> SIT\n",
        "    Vectors[7] = torch.logical_and(Agents[:,15]==3,Vectors[0]==True)\n",
        "    Agents[Vectors[7].bool(),15] = 2 # AIN --> AIT\n",
        "    # Confirmed Cases\n",
        "    # Infected and Tested\n",
        "    Agents[:,14][torch.logical_or(Agents[:,15]==0,Agents[:,15]==2)] = 1\n",
        "    # Unconfirmed\n",
        "    Agents[:,14][torch.logical_or(Agents[:,15]==1,Agents[:,15]==3)] = 3  \n",
        "\n",
        "    # ---------------------------------------------\n",
        "    # State Transition of Infected Agents\n",
        "    # ---------------------------------------------\n",
        "\n",
        "    # Find all agents who have been infected for more than a day\n",
        "    Vectors[8] = torch.logical_and(torch.logical_and(torch.logical_or(Vectors[4]==1,Vectors[4]==3),Agents[:,16]%TimeStepsDay==0),Agents[:,16]>0)\n",
        "    #Vectors[8] = torch.logical_or(Vectors[4]==1,Vectors[4]==3)\n",
        "    # SIT=0: Is the person in state 0 a child\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==0),Agents[:,0]==0)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,0][torch.multinomial(Transition[0][:,0],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # SIT=0: Is the person in state 0 an adult\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==0),Agents[:,0]==1)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,0][torch.multinomial(Transition[1][:,0],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # SIT=0: Is the person in state 0 a senior\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==0),Agents[:,0]==2)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,0][torch.multinomial(Transition[2][:,0],Population,replacement=True)][Vectors[7].bool()].double()  \n",
        "    # SIN=1: Is the person in state 1 a child\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==1),Agents[:,0]==0)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,1][torch.multinomial(Transition[0][:,1],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # SIN=1: Is the person in state 1 an adult\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==1),Agents[:,0]==1)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,1][torch.multinomial(Transition[1][:,1],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # SIN=1: Is the person in state 1 a senior\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==1),Agents[:,0]==2)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,1][torch.multinomial(Transition[2][:,1],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # AIT=2: Is the person in state 2 a child\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==2),Agents[:,0]==0)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,2][torch.multinomial(Transition[0][:,2],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # AIT=2: Is the person in state 2 an adult\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==2),Agents[:,0]==1)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,2][torch.multinomial(Transition[1][:,2],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # AIT=2: Is the person in state 2 a senior\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==2),Agents[:,0]==2)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,2][torch.multinomial(Transition[2][:,2],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # AIN=3: Is the person in state 3 a child\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==3),Agents[:,0]==0)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,3][torch.multinomial(Transition[0][:,3],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # AIN=3: Is the person in state 3 an adult\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==3),Agents[:,0]==1)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,3][torch.multinomial(Transition[1][:,3],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # AIN=3: Is the person in state 3 a senior\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==3),Agents[:,0]==2)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,3][torch.multinomial(Transition[2][:,3],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # Recovery\n",
        "    Agents[:,14][(Vectors[4]==1) & (Agents[:,15]==6)] = 5 # Reported recoveries\n",
        "    Agents[:,14][(Vectors[4]==3) & (Agents[:,15]==7)] = 2 # Unreported recoveries\n",
        "    # Deaths\n",
        "    Agents[:,14][(Vectors[4]==1) & (Agents[:,15]==9)] = 6 # Reported recoveries\n",
        "    Agents[:,14][(Vectors[4]==3) & (Agents[:,15]==9)] = 4 # Unreported recoveries  \n",
        "    # Time Since Infection\n",
        "    Vectors[8] = torch.logical_or(Agents[:,14]==1,Agents[:,14]==3)\n",
        "    Agents[:,16][Vectors[8].bool()] = Agents[:,16][Vectors[8].bool()]+1\n",
        "    Agents[:,17][Vectors[8].bool()] = torch.round(torch.div(Agents[:,16][Vectors[8].bool()],TimeStepsDay-1)) \n",
        "\n",
        "    # -----------------------------------------\n",
        "    # State Transition of Uninfected Agents\n",
        "    # -----------------------------------------  \n",
        "\n",
        "    # Find agents in the vicinity\n",
        "    CurrentPosition_numpy = np.float32(CurrentPosition[0].cpu().detach().numpy())\n",
        "    # Create a faiss array\n",
        "    index_flat = faiss.IndexFlatL2(CurrentPosition_numpy.shape[1])\n",
        "    gpu_index_flat = faiss.index_cpu_to_gpu(res, 0, index_flat)\n",
        "    gpu_index_flat.add(CurrentPosition_numpy)                  \n",
        "    m, n = gpu_index_flat.search(CurrentPosition_numpy, 20)\n",
        "    Matrixs[1] = torch.from_numpy(n).long()\n",
        "    #Matrixs[1] = torch.reshape(q[q!=torch.arange(0,Population,step=1)[:,None]],(Population,19))\n",
        "    # Distances \n",
        "    Matrixs[0] = torch.squeeze(torch.cdist(CurrentPosition[0][:,None],CurrentPosition[0][Matrixs[1].long()],p=2))\n",
        "    Matrixs[0][Matrixs[1]==torch.arange(0,Population,step=1).cuda()[:,None]] = float(\"Inf\")\n",
        "    # Location Matching and Contact Number Check\n",
        "    Matrixs[2],Matrixs[4],Contacts[0],Interactions[2],Time_Daily = generateMatrix(Agents[:,8],Contacts[0],Contacts[1],Population,H_Number,Matrixs[1].long(),Location_one, Location_two,Time_Daily,Interactions[2],Matrixs[0],Radii[1])\n",
        "\n",
        "    if time_of_day==0:\n",
        "      # Over weekends\n",
        "      if torch.logical_or(day_of_week==5,day_of_week==6):\n",
        "        Time_avg[1] += torch.mean(Time_Daily.double(),0)\n",
        "        Time_Daily = torch.zeros((Population,6),dtype=torch.long,device='cuda')\n",
        "        Interactions[1] += Interactions[2]/48#torch.reshape(torch.sum(Contacts[0],0),(6,6))\n",
        "        Interactions[2] = torch.zeros((6,6),dtype=torch.double,device='cuda')\n",
        "        Contacts[0] = torch.zeros((Population,36),dtype=torch.long,device='cuda')\n",
        "      else:\n",
        "      # Over weekdays\n",
        "        Time_avg[0] += torch.mean(Time_Daily.double(),0)\n",
        "        Time_Daily = torch.zeros((Population,6),dtype=torch.long,device='cuda')\n",
        "        Interactions[0] += Interactions[2]/48#torch.reshape(torch.sum(Contacts[0],0),(6,6))\n",
        "        Interactions[2] = torch.zeros((6,6),dtype=torch.double,device='cuda')\n",
        "        Contacts[0] = torch.zeros((Population,36),dtype=torch.long,device='cuda')\n",
        "  \n",
        "    # Find agents with infections in the vicinity\n",
        "    Vectors[3] = torch.sum(torch.logical_and(torch.logical_and(Matrixs[2].bool()==True,Matrixs[0]<Radii[1]),torch.logical_or(Vectors[4][Matrixs[1].long()]==1,Vectors[4][Matrixs[1].long()]==3)),1)\n",
        "    # Find all uninfected agents with infections in their vicinity\n",
        "    Vectors[8] = torch.logical_and(Vectors[4]==0,Vectors[3].long()>0)\n",
        "    # SUT=4: Is the susceptible person a child\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==4),Agents[:,0]==0)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,4][torch.multinomial(Transition[0][:,4],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # SUT=4: Is the susceptible person an adult\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==4),Agents[:,0]==1)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,4][torch.multinomial(Transition[1][:,4],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # SUT=4: Is the susceptible person a senior\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==4),Agents[:,0]==2)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,4][torch.multinomial(Transition[2][:,4],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # SUN=5: Is the susceptible person a child\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==5),Agents[:,0]==0)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,5][torch.multinomial(Transition[0][:,5],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # SUN=5: Is the susceptible person an adult\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==5),Agents[:,0]==1)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,5][torch.multinomial(Transition[1][:,5],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # SUN=5: Is the susceptible person a senior\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==5),Agents[:,0]==2)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,5][torch.multinomial(Transition[2][:,5],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # AUT=6: Is the susceptible person a child\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==6),Agents[:,0]==0)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,6][torch.multinomial(Transition[0][:,6],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # AUT=6: Is the susceptible person an adult\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==6),Agents[:,0]==1)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,6][torch.multinomial(Transition[1][:,6],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # AUT=6: Is the susceptible person a senior\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==6),Agents[:,0]==2)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,6][torch.multinomial(Transition[2][:,6],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # AUN=7: Is the susceptible person a child\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==7),Agents[:,0]==0)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,7][torch.multinomial(Transition[0][:,7],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # AUN=7: Is the susceptible person an adult\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==7),Agents[:,0]==1)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,7][torch.multinomial(Transition[1][:,7],Population,replacement=True)][Vectors[7].bool()].double()\n",
        "    # AUN=7: Is the susceptible person a senior\n",
        "    Vectors[7] = torch.logical_and(torch.logical_and(Vectors[8]==1,Vectors[5]==7),Agents[:,0]==2)\n",
        "    Agents[:,15][Vectors[7].bool()] = States[:,7][torch.multinomial(Transition[2][:,7],Population,replacement=True)][Vectors[7].bool()].double() \n",
        "\n",
        "\n",
        "    # -----------------------------------------\n",
        "    # Task Allocation\n",
        "    # ----------------------------------------- \n",
        "\n",
        "    # Weekday\n",
        "    if day_of_week<5:\n",
        "      \n",
        "      # Children\n",
        "      # School going children at home\n",
        "      Vectors[8] = torch.logical_and(torch.logical_and(Agents[:,0]==0,Agents[:,5]==1), Agents[:,8]==12)  \n",
        "      # Destination for school going children: School (1), Random (4), Market (7), Stay at home (12) \n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([1,4,7,12],device='cuda')[torch.multinomial(torch.tensor([Activities[2,time_of_day],Activities[0,time_of_day],Activities[1,time_of_day],1-Activities[0,time_of_day]-Activities[1,time_of_day]-Activities[2,time_of_day]]),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Stay-at-home kids\n",
        "      Vectors[8] = torch.logical_and(Agents[:,5]==0, Agents[:,8]==12)\n",
        "      # Destination for children: Random (4), Market (7), Stay at home (12) \n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([4,7,12],device='cuda')[torch.multinomial(torch.tensor([Activities[0,time_of_day],Activities[1,time_of_day],1-Activities[0,time_of_day]-Activities[1,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Kids at school\n",
        "      Vectors[8] = Agents[:,8]==13\n",
        "      # Remain in School: Stay at school (13), Go home (0) \n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,13],device='cuda')[torch.multinomial(torch.tensor([Activities[3,time_of_day],1-Activities[3,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Kids running errands\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==0,Agents[:,8]==15)\n",
        "      # Stay there (15), Go home (0)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,15],device='cuda')[torch.multinomial(torch.tensor([0.7,0.3],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Kids at the market\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==0,Agents[:,8]==17)\n",
        "      # Stay there (17), Go home (0)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,17],device='cuda')[torch.multinomial(torch.tensor([0.7,0.3],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      \n",
        "      # Adults\n",
        "      # School going adults at home\n",
        "      Vectors[8] = torch.logical_and(torch.logical_and(Agents[:,0]==1,Agents[:,5]==1), Agents[:,8]==12)  \n",
        "      # Destination for school going adults: School (1), Random (5), Market (8), Stay at home (12) \n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([1,5,8,12],device='cuda')[torch.multinomial(torch.tensor([Activities[2,time_of_day],Activities[8,time_of_day],Activities[9,time_of_day],1-Activities[2,time_of_day]-Activities[8,time_of_day]-Activities[9,time_of_day]]),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Office going adults\n",
        "      Vectors[8] = torch.logical_and(Agents[:,5]==3, Agents[:,8]==12)\n",
        "      # Destination for office going adults: Office (2), Random (5), Market (8), Stay home (12)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([2,5,8,12],device='cuda')[torch.multinomial(torch.tensor([Activities[10,time_of_day],Activities[8,time_of_day],Activities[9,time_of_day],1-Activities[8,time_of_day]-Activities[9,time_of_day]-Activities[10,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Stay at home adults\n",
        "      Vectors[8] = torch.logical_and(Agents[:,5]==2,Agents[:,8]==12)  \n",
        "      # Destination for adults: Random (5), Market (8), Stay home (12)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([5,8,12],device='cuda')[torch.multinomial(torch.tensor([Activities[8,time_of_day],Activities[9,time_of_day],1-Activities[8,time_of_day]-Activities[9,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Nomadic adults at home\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==1,torch.logical_and(Agents[:,5]==4,Agents[:,8]==12))\n",
        "      # Adults who are nomads: Go elsewhere (10), Stay home (12)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([10,12],device='cuda')[torch.multinomial(torch.tensor([Activities[12,time_of_day],1-Activities[12,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Nomadic adults on the move\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==1,torch.logical_and(Agents[:,5]==4,Agents[:,8]==16)) \n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,10,16],device='cuda')[torch.multinomial(torch.tensor([Activities[18,time_of_day],Activities[12,time_of_day],1-Activities[18,time_of_day]-Activities[12,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Adults at the office\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==1,Agents[:,8]==14)  \n",
        "      # Stay there (14), Go home (0)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,14],device='cuda')[torch.multinomial(torch.tensor([Activities[11,time_of_day],1-Activities[11,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Adults running errands\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==1,Agents[:,8]==15)  \n",
        "      # Stay there (15), Go home (0)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,15],device='cuda')[torch.multinomial(torch.tensor([0.7,0.3],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Adults at the market\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==1,Agents[:,8]==17)  \n",
        "      # Stay there (17), Go home (0)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,17],device='cuda')[torch.multinomial(torch.tensor([0.7,0.3],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      \n",
        "      # Seniors\n",
        "      # School going seniors at home\n",
        "      Vectors[8] = torch.logical_and(torch.logical_and(Agents[:,0]==2,Agents[:,5]==1), Agents[:,8]==12)  \n",
        "      # Destination for school going seniors: School (1), Random (6), Market (9), Stay at home (12) \n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([1,6,9,12],device='cuda')[torch.multinomial(torch.tensor([Activities[2,time_of_day],Activities[8,time_of_day],Activities[9,time_of_day],1-Activities[2,time_of_day]-Activities[8,time_of_day]-Activities[9,time_of_day]]),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Office going seniors\n",
        "      Vectors[8] = torch.logical_and(Agents[:,5]==6, Agents[:,8]==12)  \n",
        "      # Destination for office going seniors: Office (3), Random (6), Market (9), Stay home (12)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([3,6,9,12],device='cuda')[torch.multinomial(torch.tensor([Activities[10,time_of_day],Activities[8,time_of_day],Activities[9,time_of_day],1-Activities[8,time_of_day]-Activities[9,time_of_day]-Activities[10,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Stay-at-home seniors\n",
        "      Vectors[8] = torch.logical_and(Agents[:,5]==5,Agents[:,8]==12)  \n",
        "      # Destination for seniors: Random (6), Market (9), Stay home (12)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([6,9,12],device='cuda')[torch.multinomial(torch.tensor([Activities[8,time_of_day],Activities[9,time_of_day],1-Activities[8,time_of_day]-Activities[9,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Nomadic Seniors at home\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==2,torch.logical_and(Agents[:,5]==7,Agents[:,8]==12))  \n",
        "      # Seniors who are nomads: Go elsewhere (11), Stay home (12)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([11,12],device='cuda')[torch.multinomial(torch.tensor([Activities[12,time_of_day],1-Activities[12,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Nomadic Seniors on the move\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==2,torch.logical_and(Agents[:,5]==7,Agents[:,8]==16))  \n",
        "      # Seniors who are nomads: Go home (0), Go elsewhere (11), Stay there (16)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,11,16],device='cuda')[torch.multinomial(torch.tensor([Activities[18,time_of_day],Activities[12,time_of_day],1-Activities[12,time_of_day]-Activities[18,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Seniors at office \n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==2,Agents[:,8]==14)  \n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,14],device='cuda')[torch.multinomial(torch.tensor([Activities[11,time_of_day],1-Activities[11,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()  \n",
        "      # Seniors running errands\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==2,Agents[:,8]==15)  \n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,15],device='cuda')[torch.multinomial(torch.tensor([0.7,0.3],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Seniors at the market\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==2,Agents[:,8]==17)  \n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,17],device='cuda')[torch.multinomial(torch.tensor([0.7,0.3],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "\n",
        "    else:\n",
        "\n",
        "      # Children\n",
        "      # School going children\n",
        "      Vectors[8] = torch.logical_and(torch.logical_and(Agents[:,0]==0,Agents[:,5]==1), Agents[:,8]==12)  \n",
        "      # Destination for school going children: School (1), Random (4), Market (7), Stay at home (12) \n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([1,4,7,12],device='cuda')[torch.multinomial(torch.tensor([Activities[6,time_of_day],Activities[4,time_of_day],Activities[5,time_of_day],1-Activities[4,time_of_day]-Activities[5,time_of_day]-Activities[6,time_of_day]]),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Stay-at-home kids\n",
        "      Vectors[8] = torch.logical_and(Agents[:,5]==0, Agents[:,8]==12)\n",
        "      # Destination for children: Random (4), Market (7), Stay at home (12) \n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([4,7,12],device='cuda')[torch.multinomial(torch.tensor([Activities[4,time_of_day],Activities[5,time_of_day],1-Activities[4,time_of_day]-Activities[5,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Kids at school\n",
        "      Vectors[8] = Agents[:,8]==13\n",
        "      # Remain in School: Stay at school (13), Go home (0) \n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,13],device='cuda')[torch.multinomial(torch.tensor([Activities[7,time_of_day],1-Activities[7,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Kids running errands\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==0,Agents[:,8]==15)\n",
        "      # Stay there (15), Go home (0)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,15],device='cuda')[torch.multinomial(torch.tensor([0.7,0.3],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Kids at the market\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==0,Agents[:,8]==17)\n",
        "      # Stay there Market (17), Go home (0)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,17],device='cuda')[torch.multinomial(torch.tensor([0.7,0.3],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      \n",
        "      # Adults\n",
        "      # School going adults at home\n",
        "      Vectors[8] = torch.logical_and(torch.logical_and(Agents[:,0]==1,Agents[:,5]==1), Agents[:,8]==12)  \n",
        "      # Destination for school going adults: School (1), Random (5), Market (8), Stay at home (12) \n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([1,5,8,12],device='cuda')[torch.multinomial(torch.tensor([Activities[6,time_of_day],Activities[13,time_of_day],Activities[14,time_of_day],1-Activities[6,time_of_day]-Activities[13,time_of_day]-Activities[14,time_of_day]]),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Office going adults \n",
        "      Vectors[8] = torch.logical_and(Agents[:,5]==3, Agents[:,8]==12)  \n",
        "      # Destination for office going adults: Office (2), Random (5), Market (8), Stay home (12)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([2,5,8,12],device='cuda')[torch.multinomial(torch.tensor([Activities[15,time_of_day],Activities[13,time_of_day],Activities[14,time_of_day],1-Activities[13,time_of_day]-Activities[14,time_of_day]-Activities[15,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Stay-at-home adults\n",
        "      Vectors[8] = torch.logical_and(Agents[:,5]==2,Agents[:,8]==12)  \n",
        "      # Destination for adults: Random (5), Market (8), Stay home (12)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([5,8,12],device='cuda')[torch.multinomial(torch.tensor([Activities[13,time_of_day],Activities[14,time_of_day],1-Activities[13,time_of_day]-Activities[14,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Nomadic adults at home\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==1,torch.logical_and(Agents[:,5]==4,Agents[:,8]==12))\n",
        "      # Adults who are nomads: Go elsewhere (10), Stay home (12)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([10,12],device='cuda')[torch.multinomial(torch.tensor([Activities[17,time_of_day],1-Activities[17,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Nomadic adults on the move\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==1,torch.logical_and(Agents[:,5]==4,Agents[:,8]==16))\n",
        "      # Adults who are nomads: Go home (0), Go elsehwere (10),Stay there (16)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,10,16],device='cuda')[torch.multinomial(torch.tensor([Activities[19,time_of_day],Activities[17,time_of_day],1-Activities[17,time_of_day]-Activities[19,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Adults running errands\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==1,Agents[:,8]==15)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,15],device='cuda')[torch.multinomial(torch.tensor([0.7,0.3],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Adults returning from market\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==1,Agents[:,8]==17)  \n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,17],device='cuda')[torch.multinomial(torch.tensor([0.7,0.3],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Adults returning from office\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==1,Agents[:,8]==14)  \n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,14],device='cuda')[torch.multinomial(torch.tensor([Activities[16,time_of_day],1-Activities[16,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      \n",
        "      # Seniors\n",
        "      # School going seniors at home\n",
        "      Vectors[8] = torch.logical_and(torch.logical_and(Agents[:,0]==2,Agents[:,5]==1), Agents[:,8]==12)  \n",
        "      # Destination for school going adults: School (1), Random (6), Market (9), Stay at home (12) \n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([1,6,9,12],device='cuda')[torch.multinomial(torch.tensor([Activities[6,time_of_day],Activities[13,time_of_day],Activities[14,time_of_day],1-Activities[6,time_of_day]-Activities[13,time_of_day]-Activities[14,time_of_day]]),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Office going seniors\n",
        "      Vectors[8] = torch.logical_and(Agents[:,5]==6, Agents[:,8]==12)  \n",
        "      # Destination for office going seniors: Office (3), Random (6), Market (9), Stay home (12)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([3,6,9,12],device='cuda')[torch.multinomial(torch.tensor([Activities[15,time_of_day],Activities[13,time_of_day],Activities[14,time_of_day],1-Activities[13,time_of_day]-Activities[14,time_of_day]-Activities[15,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Stay-at-home seniors\n",
        "      Vectors[8] = torch.logical_and(Agents[:,5]==5,Agents[:,8]==12)  \n",
        "      # Destination for seniors: Random (6), Market (9), Stay home (12)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([6,9,12],device='cuda')[torch.multinomial(torch.tensor([Activities[13,time_of_day],Activities[14,time_of_day],1-Activities[13,time_of_day]-Activities[14,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Nomadic Seniors at home\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==2,torch.logical_and(Agents[:,5]==7,Agents[:,8]==12))  \n",
        "      # Seniors who are nomads: Go elsewhere (11), Stay home (12)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([11,12],device='cuda')[torch.multinomial(torch.tensor([Activities[17,time_of_day],1-Activities[17,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Nomadic Seniors on the move\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==2,torch.logical_and(Agents[:,5]==7,Agents[:,8]==16))  \n",
        "      # Seniors who are nomads: Go home(0), Go elsewhere (11), Stay there (16)\n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,11,16],device='cuda')[torch.multinomial(torch.tensor([Activities[19,time_of_day],Activities[17,time_of_day],1-Activities[19,time_of_day]-Activities[17,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Seniors at office \n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==2,Agents[:,8]==14)  \n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,14],device='cuda')[torch.multinomial(torch.tensor([Activities[16,time_of_day],1-Activities[16,time_of_day]],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()  \n",
        "      # Seniors running errands\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==2,Agents[:,8]==15)  \n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,15],device='cuda')[torch.multinomial(torch.tensor([0.7,0.3],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "      # Seniors at the market\n",
        "      Vectors[8] = torch.logical_and(Agents[:,0]==2,Agents[:,8]==17)  \n",
        "      Agents[:,7][Vectors[8]==1] = torch.tensor([0,17],device='cuda')[torch.multinomial(torch.tensor([0.7,0.3],device='cuda'),Population,replacement=True)][Vectors[8]==1].double()\n",
        "    \n",
        "    # --------------------------------\n",
        "    # Assign Task-based Goal Positions\n",
        "    # --------------------------------\n",
        "\n",
        "    for i in range(len(School)):\n",
        "      # Goal of School Going Children\n",
        "      Vectors[8] = torch.logical_and(Nearest[:,0]==i,torch.logical_and(Agents[:,8]==12,Agents[:,7]==1))# Stay at home (12) --> Go to School (1)\n",
        "      Agents[:,9][Vectors[8]==1] = School_Positions[Vectors[8]==1,2*i]\n",
        "      Agents[:,10][Vectors[8]==1] = School_Positions[Vectors[8]==1,2*i+1]\n",
        "    for i in range(len(Office)):\n",
        "      # Goal of Office Going Adults and Seniors\n",
        "      Vectors[8] = torch.logical_and(Nearest[:,0]==i,torch.logical_and(Agents[:,8]==12,torch.logical_or(Agents[:,7]==2,Agents[:,7]==3)))# Stay at home --> Office (Adults/Seniors)\n",
        "      Agents[:,9][Vectors[8]==1] = Office_Positions[Vectors[8]==1,2*i]\n",
        "      Agents[:,10][Vectors[8]==1] = Office_Positions[Vectors[8]==1,2*i+1]\n",
        "    for i in range(len(Market)):\n",
        "      # Goal of Market Going Adults/Seniors/Children\n",
        "      Vectors[8] = torch.logical_and(Nearest[:,1]==i,torch.logical_and(Agents[:,8]==12,torch.logical_or(torch.logical_or(Agents[:,7]==7,Agents[:,7]==8),Agents[:,7]==9))) # Home --> Market\n",
        "      temp1 = torch.where(Vectors[8]==1)[0].size(0)\n",
        "      Agents[:,9][Vectors[8]==1] = Market_Positions[torch.randperm(Population)[:int(temp1)],2*i]\n",
        "      Agents[:,10][Vectors[8]==1] = Market_Positions[torch.randperm(Population)[:int(temp1)],2*i+1]\n",
        "    # Goal of Adults running mobile jobs\n",
        "    mobile = goal_assign(CurrentPosition[0][:,0],CurrentPosition[0][:,1],300,800,0.5*Layout[1],0.5*Layout[3]) #300,800\n",
        "    Vectors[8] = torch.logical_and(Agents[:,0]==1,torch.logical_and(torch.logical_or(Agents[:,8]==16,Agents[:,8]==12),Agents[:,7]==10))\n",
        "    temp1 = torch.where(Vectors[8]==1)[0].size(0)\n",
        "    Agents[:,9][Vectors[8]==1] = mobile[0][Vectors[8]==1]\n",
        "    Agents[:,10][Vectors[8]==1] = mobile[1][Vectors[8]==1]\n",
        "    # Goal of Seniors running mobile jobs\n",
        "    mobile = goal_assign(CurrentPosition[0][:,0],CurrentPosition[0][:,1],300,800,0.5*Layout[1],0.5*Layout[3]) #300,800\n",
        "    Vectors[8] = torch.logical_and(Agents[:,0]==2,torch.logical_and(torch.logical_or(Agents[:,8]==16,Agents[:,8]==12),Agents[:,7]==11))\n",
        "    temp1 = torch.where(Vectors[8]==1)[0].size(0)\n",
        "    Agents[:,9][Vectors[8]==1] = mobile[0][Vectors[8]==1]\n",
        "    Agents[:,10][Vectors[8]==1] = mobile[1][Vectors[8]==1]\n",
        "    # Goal of kids running errands\n",
        "    Vectors[8] = torch.logical_and(Agents[:,8]==12,Agents[:,7]==4)\n",
        "    mobile = goal_assign(Agents[:,1].clone(),Agents[:,2].clone(),300,500,0.5*Layout[1],0.5*Layout[3])#300,500\n",
        "    Agents[:,9][Vectors[8]==1] = mobile[0][Vectors[8]==1]\n",
        "    Agents[:,10][Vectors[8]==1] = mobile[1][Vectors[8]==1]\n",
        "    # Goal of adults running errands\n",
        "    Vectors[8] = torch.logical_and(Agents[:,8]==12,Agents[:,7]==5)\n",
        "    mobile = goal_assign(Agents[:,1].clone(),Agents[:,2].clone(),300,800,0.5*Layout[1],0.5*Layout[3])#300,800\n",
        "    Agents[:,9][Vectors[8]==1] = mobile[0][Vectors[8]==1]\n",
        "    Agents[:,10][Vectors[8]==1] = mobile[1][Vectors[8]==1]\n",
        "    # Goal of seniors running errands\n",
        "    Vectors[8] = torch.logical_and(Agents[:,8]==12,Agents[:,7]==6)\n",
        "    mobile = goal_assign(Agents[:,1].clone(),Agents[:,2].clone(),300,800,0.5*Layout[1],0.5*Layout[3])#300,800\n",
        "    Agents[:,9][Vectors[8]==1] = mobile[0][Vectors[8]==1]\n",
        "    Agents[:,10][Vectors[8]==1] = mobile[1][Vectors[8]==1] \n",
        "    # Goal of Agents Going Home\n",
        "    Vectors[8] = torch.logical_or(torch.logical_and(Agents[:,8]==12,Agents[:,7]==12),Agents[:,7]==0)\n",
        "    Agents[:,9][Vectors[8]==1] = Agents[Vectors[8]==1,1].clone().detach()\n",
        "    Agents[:,10][Vectors[8]==1] = Agents[Vectors[8]==1,2].clone().detach()\n",
        "    \n",
        "    # -----------------------------\n",
        "    # Move Towards a Goal\n",
        "    # -----------------------------\n",
        "\n",
        "    # New goal position corresponding to the assigned task\n",
        "    Destination = torch.dstack((Agents[:,9],Agents[:,10]))\n",
        "    # Distance between the current position and assigned goal\n",
        "    Distance = torch.linalg.norm(CurrentPosition[0]-Destination[0],dim=1)\n",
        "    # Goal following velocity vector\n",
        "    Velocity[0] += Destination[0]-CurrentPosition[0]\n",
        "\n",
        "    # ------------------------------\n",
        "    # Collision Avoidance \n",
        "    # ------------------------------\n",
        "\n",
        "    # Reset avoidance vector in every iteration\n",
        "    VRule = torch.zeros((Population,2),dtype=torch.double,device='cuda') \n",
        "    # Find agents inside critical radius \n",
        "    Matrixs[3] = torch.logical_and(torch.logical_and(Matrixs[0]<SD[:,None],Matrixs[0]!=0),Matrixs[4]==True)\n",
        "    # Collision Avoidance Vector\n",
        "    Difference = CurrentPosition[0][:,None]-CurrentPosition[0][Matrixs[1].long()]\n",
        "    Difference[~Matrixs[3].bool()] = 0\n",
        "    # Number of agents inside the critical radius\n",
        "    Vectors[9] = torch.sum(Matrixs[3],1)\n",
        "    # Normalizing the displacement vector\n",
        "    Difference[Matrixs[3].bool()]/=torch.square(Matrixs[0][Matrixs[3].bool()])[:,None]\n",
        "    VRule[Vectors[9]>0] = torch.div(torch.sum(Difference,1)[Vectors[9]>0],torch.sum(Matrixs[3],1)[Vectors[9]>0][:,None]).double()\n",
        "    # Total Collision avoidance vector component\n",
        "    VRule[Vectors[9]>0] -= Velocity[0][Vectors[9]>0].clone().detach()\n",
        "\n",
        "    # --------------------\n",
        "    # Update Velocity\n",
        "    # --------------------\n",
        "\n",
        "    # Update the velocity of agents\n",
        "    Vectors[8] = torch.logical_and(torch.sum(Matrixs[3],1)>0,Agents[:,8]>=12)\n",
        "    Velocity[0][Vectors[8]==1] += VRule[Vectors[8]==1].clone().detach()\n",
        "\n",
        "    # ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n",
        "    # Limit Speed  \n",
        "    # --------------------  \n",
        "\n",
        "    # Speed is MaximumVelocity when outside the outer circle\n",
        "    Velocity[0][Distance>Radii[2]] = torch.div(Velocity[0][Distance>Radii[2]],torch.norm(Velocity[0],dim=1)[Distance>Radii[2]][:,None])*VelocityParameters[1]   \n",
        "    Vectors[8] = torch.logical_and(Distance<=Radii[2],torch.linalg.norm(Velocity[0],dim=1)!=0)\n",
        "    # Adjustment of velocity according to an exponential function in the intermediate region\n",
        "    Velocity[0][Vectors[8]==1] = torch.div(Velocity[0][Vectors[8]==1],torch.norm(Velocity[0],dim=1)[Vectors[8]==1][:,None])*(1000*torch.exp(-0.009*(Radii[2]-Distance[Vectors[8]==1][:,None])))\n",
        "    Vectors[8] = torch.logical_and(Distance<=Radii[4],torch.linalg.norm(Velocity[0],dim=1)!=0)\n",
        "    # Adjustment of velocity in the innermost circle\n",
        "    Velocity[0][Vectors[8]==1] = torch.div(Velocity[0][Vectors[8]==1],torch.norm(Velocity[0],dim=1)[Vectors[8]==1][:,None])*VelocityParameters[4]\n",
        "    Vectors[8] = torch.logical_and(Distance<=Radii[4],torch.linalg.norm(Velocity[0],dim=1)==0)\n",
        "    # Minimum velocity in the innermost circle\n",
        "    Velocity[0][Vectors[8]==1] = torch.sign(Velocity[0][Vectors[8]==1])*torch.tensor([0.4,0.4],dtype=torch.double,device='cuda')\n",
        "\n",
        "    # --------------------\n",
        "    # Update Position\n",
        "    # --------------------\n",
        "\n",
        "    # Update the position of agents\n",
        "    CurrentPosition[0] += Velocity[0]\n",
        "\n",
        "    # ---------------\n",
        "    # Retention\n",
        "    # ---------------\n",
        "    \n",
        "    # Change position if an agent goes out of the box (Wrap-around)\n",
        "    CurrentPosition[0][CurrentPosition[0][:,0] > Layout[1],0] = torch.DoubleTensor(1).uniform_(Layout[1]-10,Layout[1]).cuda()\n",
        "    CurrentPosition[0][CurrentPosition[0][:,0] < Layout[0],0] = torch.DoubleTensor(1).uniform_(Layout[0],Layout[0]+10).cuda()\n",
        "    CurrentPosition[0][CurrentPosition[0][:,1] > Layout[3],1] = torch.DoubleTensor(1).uniform_(Layout[3]-10,Layout[3]).cuda()\n",
        "    CurrentPosition[0][CurrentPosition[0][:,1] < Layout[2],1] = torch.DoubleTensor(1).uniform_(Layout[2],Layout[2]+10).cuda()\n",
        "\n",
        "    # -------------------\n",
        "    # Current Task Status\n",
        "    # -------------------\n",
        "\n",
        "    # Set status of those within 10 units of their destination as reached, rest in process (9)\n",
        "    # Find distance between current location and goal\n",
        "    Vectors[10] = torch.linalg.norm(CurrentPosition[0]-Destination[0],dim=1)\n",
        "    # If distance is greater than 10, agent is moving towards the goal\n",
        "    Agents[torch.logical_and(Vectors[10]>=Radii[4],Agents[:,7]<=11),8] = Agents[torch.logical_and(Vectors[10]>=Radii[4],Agents[:,7]<=11),7]\n",
        "    # Reached goal\n",
        "    # At home\n",
        "    Agents[torch.logical_or(Agents[:,7]==12,torch.logical_and(Vectors[10]<Radii[4],Agents[:,7]==0)),8] = 12\n",
        "    # At school\n",
        "    Agents[torch.logical_or(Agents[:,7]==13,torch.logical_and(Vectors[10]<Radii[4],Agents[:,7]==1)),8] = 13\n",
        "    # Adults in office\n",
        "    Agents[torch.logical_and(Agents[:,0]==1,torch.logical_or(Agents[:,7]==14,torch.logical_and(Vectors[10]<Radii[4],Agents[:,7]==2))),8] = 14\n",
        "    # Seniors in office\n",
        "    Agents[torch.logical_and(Agents[:,0]==2,torch.logical_or(Agents[:,7]==14,torch.logical_and(Vectors[10]<Radii[4],Agents[:,7]==3))),8] = 14\n",
        "    # Kids running errands\n",
        "    Agents[torch.logical_and(Agents[:,0]==0,torch.logical_or(Agents[:,7]==15,torch.logical_and(Vectors[10]<Radii[4],Agents[:,7]==4))),8] = 15\n",
        "    # Adults running errands\n",
        "    Agents[torch.logical_and(Agents[:,0]==1,torch.logical_or(Agents[:,7]==15,torch.logical_and(Vectors[10]<Radii[4],Agents[:,7]==5))),8] = 15\n",
        "    # Seniors running errands\n",
        "    Agents[torch.logical_and(Agents[:,0]==2,torch.logical_or(Agents[:,7]==15,torch.logical_and(Vectors[10]<Radii[4],Agents[:,7]==6))),8] = 15\n",
        "    # Nomadic adults\n",
        "    Agents[torch.logical_and(Agents[:,0]==1,torch.logical_or(Agents[:,7]==16,torch.logical_and(Vectors[10]<Radii[4],Agents[:,7]==10))),8] = 16\n",
        "    # Nomadic seniors\n",
        "    Agents[torch.logical_and(Agents[:,0]==2,torch.logical_or(Agents[:,7]==16,torch.logical_and(Vectors[10]<Radii[4],Agents[:,7]==11))),8] = 16\n",
        "    # Kids at market\n",
        "    Agents[torch.logical_and(Agents[:,0]==0,torch.logical_or(Agents[:,7]==17,torch.logical_and(Vectors[10]<Radii[4],Agents[:,7]==7))),8] = 17\n",
        "    # Adults at market\n",
        "    Agents[torch.logical_and(Agents[:,0]==1,torch.logical_or(Agents[:,7]==17,torch.logical_and(Vectors[10]<Radii[4],Agents[:,7]==8))),8] = 17\n",
        "    # Seniors at market\n",
        "    Agents[torch.logical_and(Agents[:,0]==2,torch.logical_or(Agents[:,7]==17,torch.logical_and(Vectors[10]<Radii[4],Agents[:,7]==9))),8] = 17\n",
        "\n",
        "    # ---------------\n",
        "    # Compute Time \n",
        "    # --------------\n",
        "\n",
        "    # Time to infection\n",
        "    Vectors[1][Vectors[4]==0] += 1\n",
        "\n",
        "    # ----------------------------\n",
        "    # Compute Reproduction Number\n",
        "    # ----------------------------\n",
        "    # With confirmed infections in vicinity\n",
        "    Vectors[8] = torch.logical_and(Vectors[4]==0,Agents[:,14]==3)\n",
        "    Matrixs[0][torch.logical_and(torch.logical_and(Matrixs[2].bool()==True,Matrixs[0]<Radii[1]),torch.logical_or(Vectors[4][Matrixs[1].long()]==1,Vectors[4][Matrixs[1].long()]==3))==False] = float('inf')\n",
        "    if torch.where(Vectors[8]==True)[0].size()[0]>0:\n",
        "      # Find the nearest individual who infected me \n",
        "      temp = torch.min(Matrixs[0][Vectors[8]==True],1)[1]\n",
        "      for l in range(temp.size()[0]):\n",
        "        Vectors[2][(Matrixs[1].long())[temp[l]]]+=1\n",
        "\n",
        "    # ---------------------------------------------\n",
        "    # Store Values of the Current Iteration\n",
        "    # ---------------------------------------------\n",
        "    #PositionTime = torch.dstack((PositionTime,CurrentPosition))\n",
        "\n",
        "    # Store Infection status for the next iteration\n",
        "    Results = torch.dstack((Results,torch.tensor([torch.where(Agents[:,14]==0)[0].size()[0],\n",
        "                                                  torch.where(Agents[:,14]==1)[0].size()[0],\n",
        "                                                  torch.where(Agents[:,14]==2)[0].size()[0],\n",
        "                                                  torch.where(Agents[:,14]==3)[0].size()[0],\n",
        "                                                  torch.where(Agents[:,14]==4)[0].size()[0],\n",
        "                                                  torch.where(Agents[:,14]==5)[0].size()[0],\n",
        "                                                  torch.where(Agents[:,14]==6)[0].size()[0]],dtype=torch.long,device='cuda')))\n",
        "\n",
        "    ResultsC = torch.dstack((ResultsC,torch.tensor([torch.where(Agents[Children,14]==0)[0].size()[0],\n",
        "                                                    torch.where(Agents[Children,14]==1)[0].size()[0],\n",
        "                                                    torch.where(Agents[Children,14]==2)[0].size()[0],\n",
        "                                                    torch.where(Agents[Children,14]==3)[0].size()[0],\n",
        "                                                    torch.where(Agents[Children,14]==4)[0].size()[0],\n",
        "                                                    torch.where(Agents[Children,14]==5)[0].size()[0],\n",
        "                                                    torch.where(Agents[Children,14]==6)[0].size()[0]],dtype=torch.long,device='cuda')))\n",
        "    \n",
        "    ResultsA = torch.dstack((ResultsA,torch.tensor([torch.where(Agents[Adults,14]==0)[0].size()[0],\n",
        "                                                    torch.where(Agents[Adults,14]==1)[0].size()[0],\n",
        "                                                    torch.where(Agents[Adults,14]==2)[0].size()[0],\n",
        "                                                    torch.where(Agents[Adults,14]==3)[0].size()[0],\n",
        "                                                    torch.where(Agents[Adults,14]==4)[0].size()[0],\n",
        "                                                    torch.where(Agents[Adults,14]==5)[0].size()[0],\n",
        "                                                    torch.where(Agents[Adults,14]==6)[0].size()[0]],dtype=torch.long,device='cuda')))\n",
        "    \n",
        "    ResultsS = torch.dstack((ResultsS,torch.tensor([torch.where(Agents[Seniors,14]==0)[0].size()[0],\n",
        "                                                    torch.where(Agents[Seniors,14]==1)[0].size()[0],\n",
        "                                                    torch.where(Agents[Seniors,14]==2)[0].size()[0],\n",
        "                                                    torch.where(Agents[Seniors,14]==3)[0].size()[0],\n",
        "                                                    torch.where(Agents[Seniors,14]==4)[0].size()[0],\n",
        "                                                    torch.where(Agents[Seniors,14]==5)[0].size()[0],\n",
        "                                                    torch.where(Agents[Seniors,14]==6)[0].size()[0]],dtype=torch.long,device='cuda')))\n",
        "  ResultsTMC[mc][0:7][:] = Results[0] # Disease\n",
        "  ResultsTMC[mc][7:14][:] = ResultsC[0] # Children\n",
        "  ResultsTMC[mc][14:21][:] = ResultsA[0] # Adults\n",
        "  ResultsTMC[mc][21:28][:] = ResultsS[0] # Seniors\n",
        "  # Overall: Time to Infection, Time to Recover, Time to Death\n",
        "  ResultsCMC[mc][0:3] = torch.tensor([[torch.mean(Vectors[1][Vectors[1]>0])/TimeStepsDay.item()],\n",
        "                                    [torch.mean(Agents[:,17][torch.logical_or(Vectors[4]==2,Vectors[4]==5)].double())],\n",
        "                                    [torch.mean(Agents[:,17][torch.logical_or(Vectors[4]==4,Vectors[4]==6)].double())]],device='cuda',dtype=torch.double)\n",
        "  # Overall: Effective Reproduction Number\n",
        "  ResultsCMC[mc][3] = torch.mean(Vectors[2][Vectors[4]!=0])\n",
        "  # Overall: Average time of infectiousness\n",
        "  ResultsCMC[mc][4] = torch.mean(Agents[Vectors[4]!=0,16]/96)\n",
        "\n",
        "  # Overall: Time to Infection, Time to Recover, Time to Death\n",
        "  ResultsCMC[mc][5:8] = torch.tensor([[torch.mean(Vectors[1][Vectors[1]>0])/TimeStepsDay.item()],\n",
        "                                    [torch.mean(Agents[:,17][Vectors[4]==5].double())],\n",
        "                                    [torch.mean(Agents[:,17][Vectors[4]==6].double())]],device='cuda',dtype=torch.double)\n",
        "  # Overall: Effective Reproduction Number\n",
        "  ResultsCMC[mc][8] = torch.mean(Vectors[2][torch.logical_and(Vectors[4]!=0,torch.logical_or(Vectors[4]==1,((Vectors[4]==5)|(Vectors[4]==6))))])\n",
        "  # Overall: Average time of infectiousness\n",
        "  ResultsCMC[mc][9] = torch.mean(Agents[torch.logical_and(Vectors[4]!=0,torch.logical_or(Vectors[4]==1,((Vectors[4]==5)|(Vectors[4]==6)))),16]/96)\n",
        "\n",
        "  # Contact Matrix\n",
        "  # InteractionsMC[mc][0][0] = Interactions[0][0]/Population/23\n",
        "  # InteractionsMC[mc][0][1] = Interactions[0][1]/23/torch.where(Agents[:,5]==1)[0].size(0)\n",
        "  # InteractionsMC[mc][0][2] = Interactions[0][2]/23/torch.where(torch.logical_or(Agents[:,5]==3,Agents[:,5]==6)==True)[0].size(0)\n",
        "  # InteractionsMC[mc][0][3] = Interactions[0][3]/Population/23\n",
        "  # InteractionsMC[mc][0][4] = Interactions[0][4]/torch.where(torch.logical_or(Agents[:,5]==4,Agents[:,5]==7)==True)[0].size(0)/23\n",
        "  # InteractionsMC[mc][0][5] = Interactions[0][5]/Population/23\n",
        "  # InteractionsMC[mc][1][0] = Interactions[1][0]/Population/8\n",
        "  # InteractionsMC[mc][1][1] = Interactions[1][1]/8/torch.where(Agents[:,5]==1)[0].size(0)\n",
        "  # InteractionsMC[mc][1][2] = Interactions[1][2]/8/torch.where(torch.logical_or(Agents[:,5]==3,Agents[:,5]==6)==True)[0].size(0)\n",
        "  # InteractionsMC[mc][1][3] = Interactions[1][3]/Population/8\n",
        "  # InteractionsMC[mc][1][4] = Interactions[1][4]/torch.where(torch.logical_or(Agents[:,5]==4,Agents[:,5]==7)==True)[0].size(0)/8\n",
        "  # InteractionsMC[mc][1][5] = Interactions[1][5]/Population/8\n",
        "\n",
        "  InteractionsMC[mc][0][0] = Interactions[0][0]/23\n",
        "  InteractionsMC[mc][0][1] = Interactions[0][1]/23\n",
        "  InteractionsMC[mc][0][2] = Interactions[0][2]/23\n",
        "  InteractionsMC[mc][0][3] = Interactions[0][3]/23\n",
        "  InteractionsMC[mc][0][4] = Interactions[0][4]/23\n",
        "  InteractionsMC[mc][0][5] = Interactions[0][5]/23\n",
        "  InteractionsMC[mc][1][0] = Interactions[1][0]/8\n",
        "  InteractionsMC[mc][1][1] = Interactions[1][1]/8\n",
        "  InteractionsMC[mc][1][2] = Interactions[1][2]/8\n",
        "  InteractionsMC[mc][1][3] = Interactions[1][3]/8\n",
        "  InteractionsMC[mc][1][4] = Interactions[1][4]/8\n",
        "  InteractionsMC[mc][1][5] = Interactions[1][5]/8\n",
        "\n",
        "# Overall: Confirmed and Cumulative=0, Total Infections (T/N)=1,\n",
        "# Confirmed Infections=2, Confirmed Recoveries=3, Confirmed Deaths=4,\n",
        "# Total Recoveries=5, Total Deaths=6, Cumulative (T/N) = 7\n",
        "FinalTimeSeries[0][:] = torch.mean((ResultsTMC[:][:,1]+ResultsTMC[:][:,6]+ResultsTMC[:][:,5]).double(),0)\n",
        "FinalTimeSeries[1][:] = torch.mean((ResultsTMC[:][:,3]+ResultsTMC[:][:,1]).double(),0)\n",
        "FinalTimeSeries[2][:] = torch.mean(ResultsTMC[:][:,1].double(),0)\n",
        "FinalTimeSeries[3][:] = torch.mean(ResultsTMC[:][:,5].double(),0)\n",
        "FinalTimeSeries[4][:] = torch.mean(ResultsTMC[:][:,6].double(),0)\n",
        "FinalTimeSeries[5][:] = torch.mean((ResultsTMC[:][:,5]+ResultsTMC[:][:,2]).double(),0)\n",
        "FinalTimeSeries[6][:] = torch.mean((ResultsTMC[:][:,6]+ResultsTMC[:][:,4]).double(),0)\n",
        "FinalTimeSeries[7][:] = torch.mean((ResultsTMC[:][:,1]+ResultsTMC[:][:,3]+ResultsTMC[:][:,5]+ResultsTMC[:][:,2]+ResultsTMC[:][:,4]+ResultsTMC[:][:,6]).double(),0)\n",
        "FinalTimeSeries[8][:] = torch.std((ResultsTMC[:][:,1]+ResultsTMC[:][:,6]+ResultsTMC[:][:,5]).double(),0)\n",
        "FinalTimeSeries[9][:] = torch.std((ResultsTMC[:][:,1]+ResultsTMC[:][:,3]).double(),0)\n",
        "FinalTimeSeries[10][:] = torch.std(ResultsTMC[:][:,1].double(),0)\n",
        "FinalTimeSeries[11][:] = torch.std(ResultsTMC[:][:,5].double(),0)\n",
        "FinalTimeSeries[12][:] = torch.std(ResultsTMC[:][:,6].double(),0)\n",
        "FinalTimeSeries[13][:] = torch.std((ResultsTMC[:][:,2]+ResultsTMC[:][:,5]).double(),0)\n",
        "FinalTimeSeries[14][:] = torch.std((ResultsTMC[:][:,4]+ResultsTMC[:][:,6]).double(),0)\n",
        "FinalTimeSeries[15][:] = torch.std((ResultsTMC[:][:,1]+ResultsTMC[:][:,3]+ResultsTMC[:][:,2]+ResultsTMC[:][:,5]+ResultsTMC[:][:,4]+ResultsTMC[:][:,6]).double(),0)\n",
        "\n",
        "# Children: Confirmed and Cumulative=16, Total Infections (T/N)=17,\n",
        "# Confirmed Infections=18, Confirmed Recoveries=19, Confirmed Deaths=20,\n",
        "# Total Recoveries=21, Total Deaths=22, Cumulative (T/N) = 23\n",
        "FinalTimeSeries[16][:] = torch.mean((ResultsTMC[:][:,8]+ResultsTMC[:][:,12]+ResultsTMC[:][:,13]).double(),0)\n",
        "FinalTimeSeries[17][:] = torch.mean((ResultsTMC[:][:,8]+ResultsTMC[:][:,10]).double(),0)\n",
        "FinalTimeSeries[18][:] = torch.mean(ResultsTMC[:][:,8].double(),0)\n",
        "FinalTimeSeries[19][:] = torch.mean(ResultsTMC[:][:,12].double(),0)\n",
        "FinalTimeSeries[20][:] = torch.mean(ResultsTMC[:][:,13].double(),0)\n",
        "FinalTimeSeries[21][:] = torch.mean((ResultsTMC[:][:,12]+ResultsTMC[:][:,9]).double(),0)\n",
        "FinalTimeSeries[22][:] = torch.mean((ResultsTMC[:][:,13]+ResultsTMC[:][:,11]).double(),0)\n",
        "FinalTimeSeries[23][:] = torch.mean((ResultsTMC[:][:,8]+ResultsTMC[:][:,9]+ResultsTMC[:][:,10]+ResultsTMC[:][:,11]+ResultsTMC[:][:,12]+ResultsTMC[:][:,13]).double(),0)                                                                                                                              \n",
        "FinalTimeSeries[24][:] = torch.std((ResultsTMC[:][:,8]+ResultsTMC[:][:,12]+ResultsTMC[:][:,13]).double(),0)\n",
        "FinalTimeSeries[25][:] = torch.std((ResultsTMC[:][:,8]+ResultsTMC[:][:,10]).double(),0)\n",
        "FinalTimeSeries[26][:] = torch.std(ResultsTMC[:][:,8].double(),0)\n",
        "FinalTimeSeries[27][:] = torch.std(ResultsTMC[:][:,12].double(),0)\n",
        "FinalTimeSeries[28][:] = torch.std(ResultsTMC[:][:,13].double(),0)\n",
        "FinalTimeSeries[29][:] = torch.std((ResultsTMC[:][:,12]+ResultsTMC[:][:,9]).double(),0)\n",
        "FinalTimeSeries[30][:] = torch.std((ResultsTMC[:][:,13]+ResultsTMC[:][:,11]).double(),0)\n",
        "FinalTimeSeries[31][:] = torch.std((ResultsTMC[:][:,8]+ResultsTMC[:][:,9]+ResultsTMC[:][:,10]+ResultsTMC[:][:,11]+ResultsTMC[:][:,12]+ResultsTMC[:][:,13]).double(),0)\n",
        "\n",
        "# Adults: Confirmed and Cumulative=32, Total Infections (T/N)=33,\n",
        "# Confirmed Infections=34, Confirmed Recoveries=35, Confirmed Deaths=36,\n",
        "# Total Recoveries=37, Total Deaths=38, Cumulative (T/N) = 39\n",
        "FinalTimeSeries[32][:] = torch.mean((ResultsTMC[:][:,15]+ResultsTMC[:][:,19]+ResultsTMC[:][:,20]).double(),0)\n",
        "FinalTimeSeries[33][:] = torch.mean((ResultsTMC[:][:,15]+ResultsTMC[:][:,17]).double(),0)\n",
        "FinalTimeSeries[34][:] = torch.mean(ResultsTMC[:][:,15].double(),0)\n",
        "FinalTimeSeries[35][:] = torch.mean(ResultsTMC[:][:,19].double(),0)\n",
        "FinalTimeSeries[36][:] = torch.mean(ResultsTMC[:][:,20].double(),0)\n",
        "FinalTimeSeries[37][:] = torch.mean((ResultsTMC[:][:,19]+ResultsTMC[:][:,16]).double(),0)\n",
        "FinalTimeSeries[38][:] = torch.mean((ResultsTMC[:][:,20]+ResultsTMC[:][:,18]).double(),0)\n",
        "FinalTimeSeries[39][:] = torch.mean((ResultsTMC[:][:,15]+ResultsTMC[:][:,16]+ResultsTMC[:][:,17]+ResultsTMC[:][:,18]+ResultsTMC[:][:,19]+ResultsTMC[:][:,20]).double(),0)\n",
        "FinalTimeSeries[40][:] = torch.std((ResultsTMC[:][:,15]+ResultsTMC[:][:,19]+ResultsTMC[:][:,20]).double(),0)\n",
        "FinalTimeSeries[41][:] = torch.std((ResultsTMC[:][:,15]+ResultsTMC[:][:,17]).double(),0)\n",
        "FinalTimeSeries[42][:] = torch.std(ResultsTMC[:][:,15].double(),0)\n",
        "FinalTimeSeries[43][:] = torch.std(ResultsTMC[:][:,19].double(),0)\n",
        "FinalTimeSeries[44][:] = torch.std(ResultsTMC[:][:,20].double(),0)\n",
        "FinalTimeSeries[45][:] = torch.std((ResultsTMC[:][:,19]+ResultsTMC[:][:,16]).double(),0)\n",
        "FinalTimeSeries[46][:] = torch.std((ResultsTMC[:][:,20]+ResultsTMC[:][:,18]).double(),0)\n",
        "FinalTimeSeries[47][:] = torch.std((ResultsTMC[:][:,15]+ResultsTMC[:][:,16]+ResultsTMC[:][:,17]+ResultsTMC[:][:,18]+ResultsTMC[:][:,19]+ResultsTMC[:][:,20]).double(),0)\n",
        "\n",
        "# Seniors: Confirmed and Cumulative=48, Total Infections (T/N)=49,\n",
        "# Confirmed Infections=50, Confirmed Recoveries=51, Confirmed Deaths=52,\n",
        "# Total Recoveries=53, Total Deaths=54, Cumulative (T/N) = 55\n",
        "FinalTimeSeries[48][:] = torch.mean((ResultsTMC[:][:,22]+ResultsTMC[:][:,26]+ResultsTMC[:][:,27]).double(),0)\n",
        "FinalTimeSeries[49][:] = torch.mean((ResultsTMC[:][:,22]+ResultsTMC[:][:,24]).double(),0)\n",
        "FinalTimeSeries[50][:] = torch.mean(ResultsTMC[:][:,22].double(),0)\n",
        "FinalTimeSeries[51][:] = torch.mean(ResultsTMC[:][:,26].double(),0)\n",
        "FinalTimeSeries[52][:] = torch.mean(ResultsTMC[:][:,27].double(),0)\n",
        "FinalTimeSeries[53][:] = torch.mean((ResultsTMC[:][:,23]+ResultsTMC[:][:,26]).double(),0)\n",
        "FinalTimeSeries[54][:] = torch.mean((ResultsTMC[:][:,25]+ResultsTMC[:][:,27]).double(),0)\n",
        "FinalTimeSeries[55][:] = torch.mean((ResultsTMC[:][:,22]+ResultsTMC[:][:,23]+ResultsTMC[:][:,24]+ResultsTMC[:][:,25]+ResultsTMC[:][:,26]+ResultsTMC[:][:,27]).double(),0)\n",
        "FinalTimeSeries[56][:] = torch.std((ResultsTMC[:][:,22]+ResultsTMC[:][:,26]+ResultsTMC[:][:,27]).double(),0)\n",
        "FinalTimeSeries[57][:] = torch.std((ResultsTMC[:][:,22]+ResultsTMC[:][:,24]).double(),0)\n",
        "FinalTimeSeries[58][:] = torch.std(ResultsTMC[:][:,22].double(),0)\n",
        "FinalTimeSeries[59][:] = torch.std(ResultsTMC[:][:,26].double(),0)\n",
        "FinalTimeSeries[60][:] = torch.std(ResultsTMC[:][:,27].double(),0)\n",
        "FinalTimeSeries[61][:] = torch.std((ResultsTMC[:][:,23]+ResultsTMC[:][:,26]).double(),0)\n",
        "FinalTimeSeries[62][:] = torch.std((ResultsTMC[:][:,25]+ResultsTMC[:][:,27]).double(),0)\n",
        "FinalTimeSeries[63][:] = torch.std((ResultsTMC[:][:,22]+ResultsTMC[:][:,23]+ResultsTMC[:][:,24]+ResultsTMC[:][:,25]+ResultsTMC[:][:,26]+ResultsTMC[:][:,27]).double(),0)\n",
        "\n",
        "# Overall: Time to Infection=0, Time to Recover=1, Time to Death=2, Rep Number=3, Time of Infectiousness=4\n",
        "# Confirmed: Time to Infection=5, Time to Recover=6, Time to Death=7, Rep Number=8, Time of Infectiousness=9\n",
        "FinalConstants[0] = torch.mean(ResultsCMC[:][:,0][torch.where(torch.logical_or(torch.isnan(ResultsCMC[:][:,0]),ResultsCMC[:][:,0]==0)==False)].double(),0)\n",
        "FinalConstants[1] = torch.mean(ResultsCMC[:][:,1][torch.where(torch.logical_or(torch.isnan(ResultsCMC[:][:,1]),ResultsCMC[:][:,1]==0)==False)].double(),0)\n",
        "FinalConstants[2] = torch.mean(ResultsCMC[:][:,2][torch.where(torch.logical_or(torch.isnan(ResultsCMC[:][:,2]),ResultsCMC[:][:,2]==0)==False)].double(),0)\n",
        "FinalConstants[3] = torch.std(ResultsCMC[:][:,0][torch.where(torch.logical_or(torch.isnan(ResultsCMC[:][:,0]),ResultsCMC[:][:,0]==0)==False)].double(),0)\n",
        "FinalConstants[4] = torch.std(ResultsCMC[:][:,1][torch.where(torch.logical_or(torch.isnan(ResultsCMC[:][:,1]),ResultsCMC[:][:,1]==0)==False)].double(),0)\n",
        "FinalConstants[5] = torch.std(ResultsCMC[:][:,2][torch.where(torch.logical_or(torch.isnan(ResultsCMC[:][:,2]),ResultsCMC[:][:,2]==0)==False)].double(),0) \n",
        "FinalConstants[6] = torch.mean(ResultsCMC[:][:,3].double())\n",
        "FinalConstants[7] = torch.std(ResultsCMC[:][:,3].double())\n",
        "FinalConstants[8] = torch.mean(ResultsCMC[:][:,4].double())\n",
        "FinalConstants[9] = torch.std(ResultsCMC[:][:,4].double())\n",
        "FinalConstants[10] = torch.mean(ResultsCMC[:][:,5][torch.where(torch.logical_or(torch.isnan(ResultsCMC[:][:,5]),ResultsCMC[:][:,5]==0)==False)].double(),0)\n",
        "FinalConstants[11] = torch.mean(ResultsCMC[:][:,6][torch.where(torch.logical_or(torch.isnan(ResultsCMC[:][:,6]),ResultsCMC[:][:,6]==0)==False)].double(),0)\n",
        "FinalConstants[12] = torch.mean(ResultsCMC[:][:,7][torch.where(torch.logical_or(torch.isnan(ResultsCMC[:][:,7]),ResultsCMC[:][:,7]==0)==False)].double(),0)\n",
        "FinalConstants[13] = torch.std(ResultsCMC[:][:,5][torch.where(torch.logical_or(torch.isnan(ResultsCMC[:][:,5]),ResultsCMC[:][:,5]==0)==False)].double(),0)\n",
        "FinalConstants[14] = torch.std(ResultsCMC[:][:,6][torch.where(torch.logical_or(torch.isnan(ResultsCMC[:][:,6]),ResultsCMC[:][:,6]==0)==False)].double(),0)\n",
        "FinalConstants[15] = torch.std(ResultsCMC[:][:,7][torch.where(torch.logical_or(torch.isnan(ResultsCMC[:][:,7]),ResultsCMC[:][:,7]==0)==False)].double(),0) \n",
        "FinalConstants[16] = torch.mean(ResultsCMC[:][:,8].double())\n",
        "FinalConstants[17] = torch.std(ResultsCMC[:][:,8].double())\n",
        "FinalConstants[18] = torch.mean(ResultsCMC[:][:,9].double())\n",
        "FinalConstants[19] = torch.std(ResultsCMC[:][:,9].double())\n",
        "\n",
        "InteractionsFinal[0] = torch.mean(InteractionsMC[:][:,0],0)\n",
        "InteractionsFinal[1] = torch.mean(InteractionsMC[:][:,1],0)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Collecting faiss-gpu\n",
            "  Downloading faiss_gpu-1.7.1.post3-cp37-cp37m-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (90.1 MB)\n",
            "\u001b[K     |████████████████████████████████| 90.1 MB 23 kB/s \n",
            "\u001b[?25hInstalling collected packages: faiss-gpu\n",
            "Successfully installed faiss-gpu-1.7.1.post3\n",
            "Collecting celluloid\n",
            "  Downloading celluloid-0.2.0-py3-none-any.whl (5.4 kB)\n",
            "Requirement already satisfied: matplotlib in /usr/local/lib/python3.7/dist-packages (from celluloid) (3.2.2)\n",
            "Requirement already satisfied: numpy>=1.11 in /usr/local/lib/python3.7/dist-packages (from matplotlib->celluloid) (1.19.5)\n",
            "Requirement already satisfied: pyparsing!=2.0.4,!=2.1.2,!=2.1.6,>=2.0.1 in /usr/local/lib/python3.7/dist-packages (from matplotlib->celluloid) (3.0.6)\n",
            "Requirement already satisfied: python-dateutil>=2.1 in /usr/local/lib/python3.7/dist-packages (from matplotlib->celluloid) (2.8.2)\n",
            "Requirement already satisfied: cycler>=0.10 in /usr/local/lib/python3.7/dist-packages (from matplotlib->celluloid) (0.11.0)\n",
            "Requirement already satisfied: kiwisolver>=1.0.1 in /usr/local/lib/python3.7/dist-packages (from matplotlib->celluloid) (1.3.2)\n",
            "Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.7/dist-packages (from python-dateutil>=2.1->matplotlib->celluloid) (1.15.0)\n",
            "Installing collected packages: celluloid\n",
            "Successfully installed celluloid-0.2.0\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "59Rcm2gD2j64"
      },
      "source": [
        "## Plots"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "b9rm1rVO2g_n"
      },
      "source": [
        "from matplotlib import pyplot as plt\n",
        "from IPython.display import HTML\n",
        "import seaborn as sns\n",
        "from matplotlib.ticker import AutoMinorLocator, MultipleLocator\n",
        "from matplotlib.colors import ListedColormap, LinearSegmentedColormap\n",
        "from matplotlib import collections as matcoll\n",
        "from matplotlib import cm\n",
        "import numpy as np\n",
        "from scipy.stats import norm"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 528
        },
        "id": "QHJka4DDTcK9",
        "outputId": "e2d87ed4-ee28-41e6-da4b-aa8c3457f60e"
      },
      "source": [
        "fig, ax = plt.subplots(figsize=(8, 8))\n",
        "\n",
        "s = 96\n",
        "\n",
        "t=96*31\n",
        "\n",
        "x = (np.arange(0,t,step=s)/96)+1\n",
        "\n",
        "data = [0.332,0.341,0.351,0.363,0.372,0.382,0.394,0.405,0.413,0.425,0.437,0.448,\n",
        "        0.459,0.470,0.479,0.486,0.497,0.508,0.518,0.528,0.537,0.546,0.553,0.562,\n",
        "        0.573,0.583,0.600,0.610,0.620,0.627,0.639]\n",
        "\n",
        "cinf_mean = ((FinalTimeSeries[0]/Population)*100).cpu().detach().numpy()\n",
        "cinf_std = ((FinalTimeSeries[8]/Population)*100).cpu().detach().numpy()\n",
        "\n",
        "ainf_mean = ((FinalTimeSeries[2]/Population)*100).cpu().detach().numpy()\n",
        "ainf_std = ((FinalTimeSeries[10]/Population)*100).cpu().detach().numpy()\n",
        "\n",
        "recovered_mean = ((FinalTimeSeries[3]/Population)*100).cpu().detach().numpy()\n",
        "recovered_std = ((FinalTimeSeries[11]/Population)*100).cpu().detach().numpy()\n",
        "\n",
        "deceased_mean = ((FinalTimeSeries[4]/Population)*100).cpu().detach().numpy()\n",
        "deceased_std = ((FinalTimeSeries[12]/Population)*100).cpu().detach().numpy() \n",
        "\n",
        "rinf_mean = recovered_mean\n",
        "rinf_std = recovered_std\n",
        "\n",
        "dinf_mean = deceased_mean\n",
        "dinf_std = deceased_std\n",
        "\n",
        "plt.plot(x,cinf_mean[np.arange(0,t,step=s)],'x-',color='blue',linewidth=2,linestyle='-')\n",
        "ax.fill_between(x, np.clip(cinf_mean[np.arange(0,t,step=s)]-cinf_std[np.arange(0,t,step=s)],0,np.inf), np.clip(cinf_mean[np.arange(0,t,step=s)]+cinf_std[np.arange(0,t,step=s)],0,np.inf) ,alpha=0.3, facecolor='blue')\n",
        "\n",
        "plt.plot(x,ainf_mean[np.arange(0,t,step=s)],'x-',color='r',linewidth=2,linestyle='-')\n",
        "ax.fill_between(x, np.clip(ainf_mean[np.arange(0,t,step=s)]-ainf_std[np.arange(0,t,step=s)],0,np.inf), np.clip(ainf_mean[np.arange(0,t,step=s)]+ainf_std[np.arange(0,t,step=s)],0,np.inf) ,alpha=0.3, facecolor='r')\n",
        "\n",
        "plt.plot(x,rinf_mean[np.arange(0,t,step=s)],'x-',color='g',linewidth=2,linestyle='-')\n",
        "ax.fill_between(x, np.clip(rinf_mean[np.arange(0,t,step=s)]-rinf_std[np.arange(0,t,step=s)],0,np.inf), np.clip(rinf_mean[np.arange(0,t,step=s)]+rinf_std[np.arange(0,t,step=s)],0,np.inf) ,alpha=0.3, facecolor='g')\n",
        "\n",
        "plt.plot(x,dinf_mean[np.arange(0,t,step=s)],'x-',color='dimgray',linewidth=2,linestyle='-')\n",
        "ax.fill_between(x, np.clip(dinf_mean[np.arange(0,t,step=s)]-dinf_std[np.arange(0,t,step=s)],0,np.inf), np.clip(dinf_mean[np.arange(0,t,step=s)]+dinf_std[np.arange(0,t,step=s)],0,np.inf) ,alpha=0.3, facecolor='dimgray')\n",
        "\n",
        "#plt.plot(np.arange(1,32,step=1),data,'o-',color='black',linewidth=2)\n",
        "\n",
        "plt.rcParams['figure.facecolor'] = 'ivory'\n",
        "\n",
        "plt.rcParams[\"font.weight\"] = \"bold\"\n",
        "\n",
        "plt.rcParams[\"axes.labelweight\"] = \"bold\"\n",
        "\n",
        "ax.grid()\n",
        "\n",
        "ax.set_axisbelow(True)\n",
        "ax.xaxis.set_major_locator(plt.MaxNLocator(16))\n",
        "ax.tick_params(which='major', length=7, color='black')\n",
        "ax.tick_params(which='minor', length=4, color='black' )\n",
        "ax.xaxis.set_minor_locator(plt.MaxNLocator())\n",
        "\n",
        "ax.tick_params(which='both', # Options for both major and minor ticks\n",
        "               top=False, # turn off top ticks\n",
        "               left=False, # turn off left ticks\n",
        "               right=False,  # turn off right ticks\n",
        "               bottom=True) # turn off bottom ticks\n",
        "ax.set_xticks(np.arange(1,32,step=2))\n",
        "ax.set_xticklabels([\"1\",\"3\",\"5\",\"7\",\"9\",\"11\",\"13\",\"15\",\"17\",\"19\",\"21\",\"23\",\"25\",\n",
        "                    \"28\",\"29\",\"31\"])\n",
        "\n",
        "plt.xticks(fontsize=16,fontweight='bold')\n",
        "plt.yticks(np.arange(0,2.2,step=0.2),fontsize=16,fontweight='bold')\n",
        "\n",
        "plt.xlabel('Time (Days)',fontsize=16,fontweight='bold')\n",
        "\n",
        "plt.ylabel('Fraction of the total population (%)',fontsize=16,fontweight='bold')\n",
        "\n",
        "plt.legend(['Total Infections','Active Infections','Recoveries','Mortalities'],\n",
        "           fontsize=16,loc=\"upper left\",fancybox=True, \n",
        "           framealpha=0.0, shadow=False, borderpad=1)\n",
        "\n",
        "plt.title('Disease evolution in Mumbai',fontsize=16,fontweight='bold')\n",
        "\n",
        "plt.savefig('FM_2.pdf',format='pdf', bbox_inches='tight',Transparent=True)\n",
        "\n",
        "files.download('FM_2.pdf')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "\n",
              "    async function download(id, filename, size) {\n",
              "      if (!google.colab.kernel.accessAllowed) {\n",
              "        return;\n",
              "      }\n",
              "      const div = document.createElement('div');\n",
              "      const label = document.createElement('label');\n",
              "      label.textContent = `Downloading \"${filename}\": `;\n",
              "      div.appendChild(label);\n",
              "      const progress = document.createElement('progress');\n",
              "      progress.max = size;\n",
              "      div.appendChild(progress);\n",
              "      document.body.appendChild(div);\n",
              "\n",
              "      const buffers = [];\n",
              "      let downloaded = 0;\n",
              "\n",
              "      const channel = await google.colab.kernel.comms.open(id);\n",
              "      // Send a message to notify the kernel that we're ready.\n",
              "      channel.send({})\n",
              "\n",
              "      for await (const message of channel.messages) {\n",
              "        // Send a message to notify the kernel that we're ready.\n",
              "        channel.send({})\n",
              "        if (message.buffers) {\n",
              "          for (const buffer of message.buffers) {\n",
              "            buffers.push(buffer);\n",
              "            downloaded += buffer.byteLength;\n",
              "            progress.value = downloaded;\n",
              "          }\n",
              "        }\n",
              "      }\n",
              "      const blob = new Blob(buffers, {type: 'application/binary'});\n",
              "      const a = document.createElement('a');\n",
              "      a.href = window.URL.createObjectURL(blob);\n",
              "      a.download = filename;\n",
              "      div.appendChild(a);\n",
              "      a.click();\n",
              "      div.remove();\n",
              "    }\n",
              "  "
            ],
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "download(\"download_5de8b488-812f-4293-8752-c9e29c1a44cd\", \"FM_2.pdf\", 20923)"
            ],
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x576 with 1 Axes>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 362
        },
        "id": "-RT46fjN2e1m",
        "outputId": "a4e046b5-4bc6-47aa-f11d-2d4360c5b96e"
      },
      "source": [
        "fig,ax = plt.subplots(figsize=(6, 6))\n",
        "\n",
        "data1 = InteractionsFinal[0].cpu().detach().numpy()\n",
        "\n",
        "data1 = np.ma.masked_where(data1 == 0, data1)\n",
        "\n",
        "data2 = InteractionsFinal[0].cpu().detach().numpy()\n",
        "\n",
        "data1[0,0] = 4.5\n",
        "\n",
        "data2[0,0] = 4.5\n",
        "\n",
        "vmin = np.min(np.array([data1,data2]))\n",
        "\n",
        "vmax = np.max(np.array([data1,data2]))\n",
        "\n",
        "alpha = ['$Q_{1}$','$Q_{2}$','$Q_{3}$','$Q_{4}$','$Q_{5}$','$Q_{6}$']\n",
        "\n",
        "cmap = ListedColormap([\"lemonchiffon\",\"goldenrod\",\"cyan\",\"orange\",\"gainsboro\", \"black\",\"yellow\"])\n",
        "\n",
        "cax = ax.matshow(data1,interpolation='nearest',cmap=cmap, rasterized=True, vmin=vmin, vmax=vmax)\n",
        "\n",
        "cbar=fig.colorbar(cax,extend='min', label='Average Contacts')\n",
        "\n",
        "plt.gca().set_xticks([x - 0.5 for x in plt.gca().get_xticks()][1:], minor='true')\n",
        "\n",
        "plt.gca().set_yticks([y - 0.5 for y in plt.gca().get_yticks()][1:], minor='true')\n",
        "\n",
        "plt.grid(which='minor')\n",
        "\n",
        "cmap.set_bad(color='white')\n",
        "\n",
        "ax.set_xticklabels(['']+alpha,y=-0.15,fontsize=12)\n",
        "\n",
        "ax.set_yticklabels(['']+alpha,fontsize=12)\n",
        "\n",
        "cbar.ax.tick_params(labelsize=12)\n",
        "\n",
        "plt.title('Weekday daily-contacts matrix',fontweight='bold',fontsize=14)\n",
        "\n",
        "plt.savefig('IFM_2.pdf',format='pdf', bbox_inches='tight',Transparent=True)\n",
        "\n",
        "files.download('IFM_2.pdf')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "\n",
              "    async function download(id, filename, size) {\n",
              "      if (!google.colab.kernel.accessAllowed) {\n",
              "        return;\n",
              "      }\n",
              "      const div = document.createElement('div');\n",
              "      const label = document.createElement('label');\n",
              "      label.textContent = `Downloading \"${filename}\": `;\n",
              "      div.appendChild(label);\n",
              "      const progress = document.createElement('progress');\n",
              "      progress.max = size;\n",
              "      div.appendChild(progress);\n",
              "      document.body.appendChild(div);\n",
              "\n",
              "      const buffers = [];\n",
              "      let downloaded = 0;\n",
              "\n",
              "      const channel = await google.colab.kernel.comms.open(id);\n",
              "      // Send a message to notify the kernel that we're ready.\n",
              "      channel.send({})\n",
              "\n",
              "      for await (const message of channel.messages) {\n",
              "        // Send a message to notify the kernel that we're ready.\n",
              "        channel.send({})\n",
              "        if (message.buffers) {\n",
              "          for (const buffer of message.buffers) {\n",
              "            buffers.push(buffer);\n",
              "            downloaded += buffer.byteLength;\n",
              "            progress.value = downloaded;\n",
              "          }\n",
              "        }\n",
              "      }\n",
              "      const blob = new Blob(buffers, {type: 'application/binary'});\n",
              "      const a = document.createElement('a');\n",
              "      a.href = window.URL.createObjectURL(blob);\n",
              "      a.download = filename;\n",
              "      div.appendChild(a);\n",
              "      a.click();\n",
              "      div.remove();\n",
              "    }\n",
              "  "
            ],
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "download(\"download_c63a7f10-850a-4760-9480-6eebaba036f2\", \"IFM_2.pdf\", 16427)"
            ],
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXgAAAFZCAYAAACSbTQ1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3deVyU5eL+8c+AbC6IOyhuiCuopJgkarjkkmGW5Ypbx0izU576VscWK82TaZZZWWInMStN2wxLs4UxcUHRNA2RTBFMLZdQUMzt/v3BzymOCwOCw4zX+/WaV8wzz33PNdNwcfswPGMxJtsgIiIux83RAUREpHSo4EVEXJQKXkTERangRURclApeRMRFqeBFRFxUOUcHcGYNGrRk794sRowYTHz8m3aPs1pX06VLNACJiQlERXUqrYiXlZGxl4YNWwMwb94bjBw51O6x//u4r2YucR3F/X64lkaOHMvOnbuKPf4//3maLl06l2Ci0uV0BT99+iwee2wibm5uHDmyGz8/PwDuvnsEH320FIBff91B7doBAIwadT/x8R9Qr14ge/dud1huV3LDDa3w969Fo0YNHR2lRMTHv8+oUeMA2LNnKw0a1C/x+7BY8l+nzzzzOM8+O6HE5y9Jxf2B7Qyvix9/XMj69cUbu2ABbNuWqoIvTZ07dwDg/PnzrF69jujo3gCsXr3Ots/3369l0KD+AKxatabAOLl6n376vqMjSBly+vRpPD09neJ1YbGAp2fxxpZzurZ0wmPwbduGUaFCBSC/yAHS03fx22+/U7NmDeCvst+371f27NkLQOfOkQAYY3jjjbm0bh2Jj48/lSvXo2/fQaSmphW4n/T0XQwadA81awbj6VmDxo3bMH36LM6fP3/ZbLt27aZ27WZYLH7ccks/8vLyAHjrrXeoVy+U8uUDiI4eyK+/Hrho7KZNW+jWrS8BAU3x8qpJhQq1adeuC++99yEAaWnpWCx+WCx+LF/+tW3cd9+tsm1PS0u/bLZPP02gSZO2eHvXonPn3qSm7rxon717M+nd+y7q1g3Bx8cfHx9/QkNvYubM2Rjz1x88N2jQEovFj5Ejx17yvq4264Xno1+/IVSvHoSXV03q1w/lueem2m7PzMxi+PD78PdvgodHderUaU5s7EP8/vsh2z4jR47FYvGjQYOWLFnyGc2ataNChdp07tybnTt/tu1zYfUO0LBhaywWP5599gUAHn30aUJCIvDzq4eHR3Vq127GiBFjOHDgoF15rdbVttU7wHPPvWjLBPmvszvuGEqtWo3x8qpJ7drN6N79dr766tvLPjfx8e/bnsdPPvmc8PAofHz8ueWWfhw4cJAPPlhCUFBr/PzqMXToveTk5NjGvvTSa4SFdaRq1QZ4eFSnRo1G3HlnDOnpu2xzX1i9A4waNQ6LxY+oqD4AREX1sV1/8cWZ1K7djFq1GgMXvy4eeeRJLBY/qlcP4rfffgdg0qT8x+/rW5fduzOu+BqQEmBMtnG2S/fuUQYwN97Y1hiTbebOnWUAM3nyk8bDw8OEhrYwxmSb99+fawADmB07Nhhjss0DD9xr29a8eVNTq1ZNA5jKlX3NL79sMcZkm59/3mz8/CobwPj5VTatWoUYNzc3A5gHHrjXlqN+/boGMCNGDDZ7924z9eoFGsD07NnN5OUdNMZkmy++WGy7v6pVq5igoAamQoUKtm2JiQnGmGyzZMl84+bmZurXr2tuuKGVqVLFz7bPsmUfGmOyTbduNxvA3HXX7bYMY8f+wwCmffvwyz5fP/64xri7uxvAVKpUyTRt2rhAhnnz3jDGZJuNGxMNYAID65gbbmhlatasYdvn9denX/JxG5Nt9uzZetFcxc1qTLZZs+Yr4+npaQDj4eFhQkKamxo1qpubb440xmSb33772dSuHWAA4+XlZVq0aGY8PDwMYBo3bmRycvYZY7LNiBGDDWDKlStnPDw8TLNmTYzFYjGA6dChvTEm20ya9IQJCmpgyx8W1tK0bx9u5s6dZYzJNiEhzU3lyr4mNLRFgfHt2rWxK++mTVbTvn24bf46dWqb9u3DTb9+fYwx2aZNm9YGMFWq+Jk2bVqbOnVq217Ll3t+5s17wzafj49PgVzNmjUxXl5epkmTYNs+TzzxiG1snz49TYUKFUzz5k1NaGgL2+siMLCOycs7aJYt+9CEhbW0jQ0KamDatw83Y8f+wxiTbW6+OdIAxtPT07i7u5sWLZqZhg3rX/J1cerUb6ZVqxADmP79+5otW1bb/j/Fx892SHe0aYMxpniX+HjMq69OdXj/FeXilAU/adITtm/cnJx9ZtiwgQYwGzZ8Z2666UZjsVjMkSN7zH33jTKAqVmzhjEmv4gufCPMmTPTXHgRhoQ0N4AZPXq4MSbbjBo11ACmSZNgc/x4ljEm27z3XpwBjJubm8nM3G7+/oLu2bObCQ4OMoDp06enOXXqN1vWzp07GMDUr1/XZGfvNcZkm5iYARcV/P79aebgwXTbuLy8g7Y5Y2IGGGOyzSefLLB9cx069Is5d+6o8fevZQDz5psvX/b5Gj58kAFMxYoVTVbWT8aYbPPkk/93USn/8UeG2bNnq23cuXNHbfk7drzJtt2egi9uVmOyTZcunWw/dFNTk40x2eb8+T/M5s2rjDHZZuLExwxgLBaL2bDhO2NMtlm+/CNbhlmzXjTG/FXwgPn884XGmGzzr3/db9t28uQBY0zBwvz74zcm22zdmmTOnTtqu35hMQGYXbt+sCuvMdm2Mc8883iB+StWrGgA8/33X9q2ZWX9ZNLSNl72+fl73ueff8oYk22GDv3rNfXee3HGmGzTseNNF/1A3b59nTl9+pDt+tdff2Yb9803Sy/7//PC5ULBA2b58o+MMdnm7Nkjl3xdGJNttm1ba7y9vQ1gW0wNGHCHw7rjeit4pztEA38dTz979ixr125g9ep1VKxYkTZtWtO5cweMMSQlrbMdf+/YMQKAjRt/sB1quO++8Vgsfnh71+Knn3YAsH59CgDJyZuA/H8++/rWxWLxIyYmFsg/9r9hw6YCeb766lt27dpN27ZhfPLJAry8vGy3bduWCkDPnt2oXLkyAAMG3HHRY3Jzc+ORR56idu1mlCtXDR8ff3bt2g3A/v35hwP69r2VunUDOX36NO+99yFr1qzn4MHf8PLyYtCgOy/7fF3I0KHDjQQG1gFg4MCLM3h4eDBt2izq1w/Fw6M67u5VbYfB9u+/+LDSldiTdfLkaUREdLddJk+eBvz1/N9xx200b94UAIvFwg035B862LjxBwCCg4No164NAL16dadKlfxDISkpPxTIUrmyr+13NS1aNLNt//vhnMvZunU77dp1oWLFOlgsftx774O22y48J4XlvZLo6F4AdOt2O02bhnP77YNZsuQz6tQJKHTs38c3aFDvom1BQfm/LL5weAQgM3MfXbpE4+tbFze3KtxyS7+LHo89mjZtTK9e3QFwd3e/7H6hoS2YOvUZW446dWozZ84rdt+PXB0n/LUBtG8fjpeXF3/++ScffLCEjIxMevToiru7O507d+DFF2fy8cef247zXuoXrK1bh+Lt7V1gW+3a/gWuV6tWleDgoIvG+vj4FLhesWJFcnNz2bJlGwkJy+nf//YiP6aYmFi++caKxWKhRYtmVKxYgdTUneTk5HDu3Dkg/xspNnYETz89hXfeeY+MjEwAbr/9Vtu7ia7G+PETePvtdwFo3LgRVatW4Zdf9nD48BFbBnvZk/WXX/aQnJxiG9OsWeOrfgyX4udX2fZ1uXJ/ldGFH/aXk5S0jhEjxmKMoVq1qrRo0ZTc3BPs2JH/+4uiPieX8u67b9G3b2+s1iRSU9NYuTKRzz9fjtWaxNKlCwsd7+tbCSj4uHx9fYH8HzLw1+PcvTuDfv2Gcvr0aSpVqkTbtmGcPXuWLVu2Ffnx1KpVw+59L/y/B8jOPsbhw0dL5PUqhXPKFby3t7dt5fb++0uAv0q8Y8cI3Nzc+OCDj2z7X7gtPDzM9qIfPPgu1q//xnZ5442X+L//+ycA7drdAECFCuVJSFhk22flyk8YO/Yebr21R4E8/ftHM3z4IM6dO8eQIfeycuV3tttCQ5sDsHLld7Zfdl14O+ffXfjXw733jmD79nV8+eUSKlascNF+9947Ak9PT7ZtS+XttxcAMGLE4Cs+XxcyrF27wbZKW7Lks8tm6NGjK+npm7Bal9m9kryUwrLGx7+JMdm2y4X3Trdv3xaAzz77wvbLP4CtW/OL6ML/n127drNx42YAVqz4hj/+yAYgPPyGIuUsX7687esTJ07avk5OTrGV47Zta9mw4TuGDx900fjC8sJfi4K/zw+wevVa7rjjNt566xW+/345zzzzOADffbe6SI/BHj/8sJXTp08D8NVXH7NxYyKPPz7+ov0u93z83YXvo8J8/XUir776Fm5ubrRqFcKJEyeIiYnl7NmzxXgEUlROWfBQ8DDN36/7+vrSunWobXv+9fx3LDRs2IAxY+4B4N//fpb69UNp3TqSqlUbEB4eZSvmJ554hMqVfcnM3Ef9+i0JC+tIw4atqFYtiJEj779knrfffo2ePbtx+vRp7rxzGOvWbQDgscceAvJXMQ0btqZRozA+/PDTi8a3ahXy/+d5l5CQCBo1CuPUqVMX7VerVk369+8LwIkTJ/D3r0XPnt2u+Fw98sgDuLm5kZubS7NmN9K8+Y1MmzbrshlWrvyOpk3DqVs3hKysX68495UUJyvA888/haenJ9nZxwgNvYmWLTvg79+Ehx76NwDjxt1LQIA/xhg6depNaOhN9O2b/4MjODiIUaOK9odWf/+XQ/fu/YiI6M6aNetp1SrUtr1lyw40b34j06df/LwVlvfv9zFr1hzatevCE09MAmDYsDFUqdKApk3DueGGTkyc+B/gr/8XJSkkpLntcEqvXnfRsmUH/vnPxy7ar0aN6lSrVhWAf//7Odq378Zrr80p8v0dOXKUkSPvxxjDQw+N4YsvFuPnV5nk5BTb4TgpXU5f8JC/or/xxra26zffHGn7OjKyPW5ufz3M11+fzqxZL9K6dSi//36YPXsyCQjwZ+zYf9jKqEmTYJKTv2XQoP5UqlSR1NSdnD59hqiojsyc+cIl83h4ePDxx+8SHn4DJ06coE+fAWzb9hO33daL11+fTp06tTlx4iSNGjXkzTdnXDQ+Pn42Xbp0wtvbm5MnTzJz5guX/SYfN2607euYmAFXPAYK0Lp1Sz78cB7BwUH8+eefVK7sy8KFb1+038svT+H222+lYsWK5OTk8uijD9qO5xZXUbMCdOjQnrVrV3L77bfi61uJnTt/xsvLk6iojgDUrFmD9eu/Ztiwgfj5VWbnzp+pUaM6o0cPZ82ar6hYsWKRMrZqFcrTTz9KrVo1OXjwN5KTU/jjj2xuuaULL774HLVrB5CXd4pmzRrz5psvFzkvwKxZL9KyZQsg/3cEF1b699wzlJYtW3DkyFF++imNGjWqM3ToABYt+m+RHoM9mjVrwjvvvE7DhvU5ffo01atXZeHCi+/HYrEwd+6rBAcHkZeXx4YNm9i7N6vI9xcb+xD79x+gSZNgpkx5msDAOrz2Wn6xT5kyw7YIktJj+f+/4RcncujQYfz9m3D+/Hl++ml9gV8cljXOlFVcX9u2fmzaVPh+lzJ/Phw7NpUHHxxTsqFKkdOu4K9Hhw8fYejQe+nUqTfnz5+nT5+eZbYwnSmriKtyynfRXK9yc3P54IMleHt706NHV/7739ccHemynCmriKtSwTuRBg3qY0y2o2PYxZmyirgqHaIREXFRKngRERelghcRcVEqeBERF6WCFxFxUSp4EREXpYIXEXFRKngRERelghcRcVEqeBERF6WCFxFxUSp4EREX5bIFHxcX7+gIRabMpc/Z8oLzZXa2vK7M4QWfkLC8VOYtzReZs2UurbzgfJmdLS84X2Zn/N5zVWWg4Fc4OkKROVtmZ8sLzpfZ2fKCMl8PHPaRfQkJy0lIWMGCBR9SvV69kr+Ds2ehXOmc7t7z7FnKlcLcZ51s3tKc++yZM5TzKPzzW4s879lzlCtXCvOeOVcqeaEUM58973SvtwMH9nP8+L5ij7/ePrLPYR/4ER3dm+jo3ixftY59O3c6KkaxJFqtREVFOTqG3axOlhfAmvglUR39HR3Dbtakg06VF8C65rjTvS7Cw29wdASn4vBDNCIiUjpU8CIiLkoFLyLiolTwIiIuSgUvIuKiVPAiIi5KBS8i4qJU8CIiJWTmzNm0atUBd/eqWCx+PPvsCwVuX7LkM0JCIvDyqkmDBi2ZNu3VUs2jghcRKSGbNm2hatUq1K1b56Lb1q3bwMCBo8jM3MegQXdy9uw5Hn/8GebMmVdqeVTwIiIlZMGCOKzWLwgLa3nRbS++OBNjDM8++zjz57/F/PlvAvDCCy+XWh4VvIiIndauTSY8PMp2KcqZM3/44Ufgr9MthIeHAbB3bxbZ2dklnhUceC4aERFn06FDexYtKt4hld9+OwRAxYoVAahQoYLttoMHf8fPz+/qA/4PreBFRK6BWrVqAJCbm1vgvwD+/jVL5T5V8CJy3di8GSyW4l1Gjry6+75wXH7Dhs0AbNz4AwD16gWWyuoddIhGRKTEvP32uyQlrWPz5vzj7Z999gUZGZn069eHxx57iISEFTz33Its357KN9+sAuDf//5XqeXRCl5EpIQkJa1j/vyFZGXlfyjJ1q3bmT9/IVu2bCMyMoKFC/9LvXqBLFz4Me7u7rzwwjOMGXNPqeXRCl5EpITEx79JfPybl7194MA7GTjwzmuWx+4VfE5ODhMnTiEkJILy5QMICGhKv35DSE1NK818IiJSTHYV/P79BwgP78LkydMxxjBs2EBq1KjG0qVf0q5dV7Zt+6m0c4qISBHZVfAxMbGkp++iR4+ubN2axJw5M9m0aRVBQQ04efIkr7wyu7RziohIERVa8MnJKSQmrgZgxozn8fDwAMDDw4OIiHYAZGYW7VPO4+LibX8JduLYsaJmFhEROxRa8FZrEgDVq1cjNLRFgdsOHToMQJ06AZw5c4aIiO54eFTH37/JFeeMjR1JSoqVlBQrFSpXLm52ERG5gkLfRXPkyFEAqlatUmB7Tk4OycmbAIiMjMBisXDHHbdRsWIFtm/fUQpRRUSkKApdwQcHBwGwZ89e23s7ASZPns7x48cJCPBn2LCBlCtXjscfH09gYO3SSysiInYrdAU/eHB/XnjhZTIyMomM7EWvXt3YsSOdpKR1+Pr6smzZInx8fK5FVhERKYJCV/CVKlVi9erlxMQMIC8vj7lz55OUtI6uXTuTlraBNm3CrkVOEREpIrveJhkYWIcFC+I4dOgXUlKseHl5kZq6E3d39wL7paWlc+zYcc6dO0daWjonT54sldAiIlK4Ip+qoG3bMBYvnsfmzVvZsmUbPXp0td3WvPmNBb5OTEwgKqpTySQVEZEiKda5aPr2vZW+fW+9aLsxpfOpJCIiUnQ6m6SIiItSwYuIuCgVvIiIi1LBi4i4KBW8iIiLUsGLiLgoFbyIiItSwYuIuCgVvIiIi1LBi4i4KBW8iIiLUsGLiLgoFbyIiIsq1tkkS0JCwnISElZwJieHl6xWR8UoltzcXKxOlNnZ8gLk5p7FmnTQ0THs5mx5AXJPnHe614UUjcMKPjq6N9HRvVnV9EYeiYpyVIxisVqtRDlRZmfLC86X2dnygnNmlqLRIRoRERelghcRcVEqeBERF6WCFxFxUSp4EREXpYIXEXFRKngRERelghcRcVEqeBERF6WCFxFxUSp4EREXpYIXEXFRKngRERelghcRcVEOO12wiMi1FhISwtKlS4s19pNPPgHOl2ygUqYVvIiIi1LBi4i4qCIVfE5ODhMnTiEkJILy5QMICGhKv35DSE1NK618IiJOZcuWH+nZ806qVWtI+fIBtGjRntmz33ZIFrsLfv/+A4SHd2Hy5OkYYxg2bCA1alRj6dIvadeuK9u2/VSaOUVEnEK/fkNZufI7GjVqSP/+0aSlpTNu3P+RmPj9Nc9id8HHxMSSnr6LHj26snVrEnPmzGTTplUEBTXg5MmTvPLK7NLMKSJS5p05c4asrF8BeOed11mwII42bVoDkJGRec3z2FXwyckpJCauBmDGjOfx8PAAwMPDg4iIdgBkZu4rpYgiImXD2rXJhIdH2S5xcfEFbvfw8OChh8YAcM89DzBsWCybN2+ldetQ7rjjtmue166Ct1qTAKhevRqhoS0K3Hbo0GEA6tQJwGpdTdOm4Xh716Jx4zZ89dW3l5wvLi7e9gQdO3bsavKLiFwzHTq0JyXFarvExo68aJ9+/frQoEE9Nm7czHvvLaZcuXL069eHSpUqXfO8dhX8kSNHAahatUqB7Tk5OSQnbwIgMjKCU6f+ZMKEf7F58yoqV/blvvvGX3K+2NiRtieocuXKV5NfRKTMOHLkKL17301GRiarVy/n6NEMwsJa8txzLzJnzrxrnseuP3QKDg4CYM+evWRl7aNu3UAAJk+ezvHjxwkI8GfYsIH4+PjYxrRtG8bChR+XQmQRkbJpz569nDx5Eg8PD9q1a4OXlxfNmzdh48bN7Nix85rnsavgBw/uzwsvvExGRiaRkb3o1asbO3akk5S0Dl9fX5YtW1Sg3LdvT+X995cwZsyoUgsuIlLWNG/ehKpVq3D06B9063Y7jRo1sC10O3a86ZrnsesQTaVKlVi9ejkxMQPIy8tj7tz5JCWto2vXzqSlbaBNmzDbvj///Au33HIHnTrdxNSpz5ZWbhGRMqdChQp8+eUSunePYseOnSxZspTg4CBmznyBgQPvvOZ57D4XTWBgHRYsiANg06YtREb2JDV1J+7u7rZ9srL20b377dSsWZ1Zs17k4MHfCAjwL7CPiIgra98+nK+//szRMYBinqqgbdswFi+ex333jWTLlm227d9+u4rMzH38+ONPNGnSlrp1Q8jK0tsnRUQcodhnk+zb91b69r21wLaRI4cycuTQqw4lIiJXTycbExFxUSp4EREXpYIXEXFRKngRERelghcRcVEqeBERF6WCFxFxUSp4EREXpYIXEXFRKngRERelghcRcVEqeBGRMsxqXc0331g5d+4cEydOYfTof/Lzz7/YNbbYJxsTEZHS98ADj3H77bdy6NBhnn/+JQDS03fx/ffLCx2rFbyISBm2e3cGzZo1Zu3aZAYN6s8rr/yHzZt/tGusw1bwCQnLSUhYQU5ODlar1VExiiU3N9epMh8/fpyvvvrK0TGK5MzpP7EmfunoGHbLzT3rVHkBck+cd6rX8fXKx8ebL75YyZYt27j//n9QoUJ5uz9EyWEFHx3dm+jo3jRteiNRUVGOilEsVqvVqTJ/9dVXBAcHOzpGkWRl7iSqo7+jY9jNmnTQqfICWNccd6rX8fXq7rv7ERcXT4UKFejXrw8vvvgqLVo0tWusjsGLiJRhb731CmPGjKJ27QBq1qzB+PFj8fHxtmusjsGLiJRh7u5V+fnn3dSsWQOAn3/+ha5d+9o1Vit4EZEyKDMzi4yMTIwx/PTTDmrVyi/45cu/ZvfuDLvmUMGLiJRB8+a9z6RJ07BYLEyePJ3Jk6cDYIyheXMdgxcRcVo33tiWsWP/wezZb9OjR1caNw7CYrFQpYofQ4cOsGsOFbyISBnUu/ct9O59C+3a3UBUVEfq169X5Dn0S1YRkTLs229XMW/e+7brzzzzH4YPv8+usSp4EZEy7OOPE6hfv67tev36dfnkk2V2jVXBi4iUYX5+lVm1ao3tutWaROXKvnaN1TF4EZEyLDq6F3Fx8Xz11XcA/P77IWJjR9o1VgUvIlKGTZ8+idOnT7NsWf75pEaOHMK0ac/ZNVYFLyJShlWqVIl33nmjWGNV8CIiZdiBAwd59tmpbNuWyqlTpwCwWCxs2rSq0LEqeBG5bnid+IlGycU7s2rNX+BYi6klnKhwo0c/yFdffYsxhnLlynHmzBn8/CrbNVbvohERKcPWrk1mwoR/AfDFF4sZM+Yexo0bbddYFbyISBl2+vQZGjasjzGGLVu2UalSRV59dY5dY3WIRkSkDGvQoB6HDx+lVasQHn/8GQCaNWti19gireBzcnKYOHEKISERlC8fQEBAU/r1G0JqalrRU4uISKEWL57HHXfcxnvvxXH33f0YMOAOFi+eZ9dYuwt+//4DhId3YfLk6RhjGDZsIDVqVGPp0i9p164r27b9VOwHICIilzZjxutkZx8jNLQFixa9w8SJj7F0qX2f/2t3wcfExJKevosePbqydWsSc+bMZNOmVQQFNeDkyZO88srsYj8AERG5tPj4D8jIyLRd37hxM08/PcWusXYVfHJyComJqwGYMeN5PDw8APDw8CAioh0AmZn77A4cFxdPeHgU4eFRHDt2zO5xIiLXi1dffZOgoNZYLBYeeOBRgoJaExTUmnHjHi3Zt0larUkAVK9ejdDQFgVuO3ToMAB16gSQmppGgwYt8fKqSVBQaz7+eOkl54uNHUlKipWUFCuVK9sXVETkenLyZB6HDh0B4PjxHA4dOsLhw0epWtWPSZOesGsOuwr+yJGjAFStWqXA9pycHJKTNwEQGRmBr28l3n77NbZsWU3FihV48snn7X4wIiLylwkTHiYnZx+dO3dg+fIl5OTs4/jxLPbu3c64cffaNYddb5MMDg4CYM+evWRl7aNu3UAAJk+ezvHjxwkI8GfYsIH4+PgQGFiHM2fO4OvrS40a1Yr50EREBCAxcRlr1qzn/fcXc+7cOdv24cMHFzrWroIfPLg/L7zwMhkZmURG9qJXr27s2JFOUtI6fH19WbZsET4+PgDMnv0248dPAGDevOKdIEdERPLFxMSycOFHtuvGGCwWS8kVfKVKlVi9ejkTJjzHihXfMnfufAC6du3Me+/FERDgb9t36NC76dgxggcffJyxYx9hyJC7sVgsRX1MIiICJCSsoG3bMPr370u5cu5FGmv3X7IGBtZhwYI4ADZt2kJkZE9SU3fi7v7XHa5Zsx53d3eqVq2Cp6cn3t5eKncRue58+mkC//nPy2zfvgNPT09atmxBQsIiqlTxK/JckZHtufnmSB5/fHyRxxbrVAVt24axeGLjeA8AABuCSURBVPE8Nm/eypYt2+jRoysAu3dn8MgjT3H8eA7BwUHMn/9mcaYXEXFaCxd+xJAho/Hy8qJfvz5UrFiBDRs2cfLkyWIV/LFjx3nqqedZtuwrqlTJf9ehxWJh6dKFhY4t9rlo+va9lb59by2wbdiwQQwbNqi4U4qIODVjDI8//iwAK1Z8RFRUp6uec926DUD+EZIL7D0yopONiYjYae3aZN59d5HtemzsyAKfj/rzz7+QlbUPHx8fpk2bxW23DcLfvyb/+tf9dr+18X/t2bO12HlV8CIidurQoT2LFl3+RF+HD+f/YVJeXh67d2cwYEA/Fi78mAceeJQ6dQLo1++2It9n/fr1sFpXs3z5NwD06dODzp0j7Rqr88GLiJSQGjWq275esGAO77zzBvfcEwPA558vL9acc+fOp1u325k+fRbTp8+iS5do/vvfd+0aq4IXESkh9evXxdfX95K3VaxYoVhzTpv2KhER7Vi58lNWrvyUm266kRdffNWusSp4EZES4unpyfjxYwAYPnwM99wzjnfeeQ93d3eGDh1QrDkPHvydESMG0717FN27RzF8+CAOHPjNrrE6Bi8iUoKefvoxTp8+Q3z8B3z44aeEhjbnuecm0L59eLHma9GiKc8+O5V9+34F4L//fY/Q0OZ2jVXBi4iUoHLlyvHCC8/wwgvPlMh8M2Y8T9++g3n++ZcAqFLFjxkz7DuRowpeRKQM69jxJnbt+sH2fvibbrrxojP7Xo6OwYuIlEGffbaMp5/OX6lXrVqFPn16cuutPZg5czaffbbMrjlU8CIiZdDEif/h5Mm8AtssFgt5ead46qkS/Mg+ERG5tnbt2kNISLOLtjdv3oTduzPsmkMFLyJSBvn6VuLbb7+/aPu3335PpUoV7ZpDv2QVESmDevfuzrvvLiI7+xg9e+afsXflykSWL/+aESMK/7APUMGLiJRJ06ZNYv36FJYv/5oVK/LPQ2OMoVmzJkybNsmuORxW8AkJy0lIWEFOzjGs1hmOilEsublNsVqtjo5htzOn/yQrc6ejYxRJbu5ZrEkHHR3Dbs6WFyD3xHmneh1fb2rUqM4PP3zPBx8sYdOmrVgsFtq2bc3gwXfh7e1t1xwOK/jo6N5ER/emadNQoqL+z1ExisVqTSQqKsrRMexmTfySqI7+he9YhliTDjpVZmfLC2Bdc9ypXsfXI29vb+65Zxj33DOsWOP1S1YRERelghcRcVEqeBERF6WCFxEpw/Ly8nj00acJC+vImjXrefDBx1i8+FO7xqrgRUTKsPHjJ/DKK7PZti2VP//8k3PnzjN9+iy7xqrgRUTKsE8+SeDRRx+0XW/btjU7d+6ya6wKXkSkDHNzc8MYY7u+det2uz/+T3/JKiJShvXp04OXX34DgGHDxnDw4G+MHj3crrEqeBGRMmzmzBcwxvDFFys5c+YMI0YMZvr0Mn6qAhERKZyvry/z5s0u1lgVvIhIGda1a/RF2/z8KnPLLV0YO/YfVxyrghcRKcOs1qRLbl+69EsOHz7C008/dtmxeheNiEgZ9uSTjxAd3Yv09E3s3JlCdHQv/vWv+xky5C7mz194xbFawYvIdWNz1TZYhmwq3uAz83n12LU/JfQbb7zNhAkPExwcBECnTh2YMeN13nsvjo8++vyKY1XwIiJlWJ06tXnyyckkJCzHYrGwbt1GmjdvwpEjR6lWreoVx6rgRUTKsA8+mMuIEWNJSloPwA03tCI+fjZHj/7BrFlTrzjW7oLPyclh+vRZfPxxAnv27KVyZV/at2/Lf/4zkRYtLv7kbxERuXotW4awefP3HD9+HMh/26S97Cr4/fsP0KVLNOnpu2jevCnDhg1k3boNLF36JV9/bWX9+q9p2TKkeOlFROSyjDEsWvQx27alcurUKQAsFgszZkwpdKxdBR8TE0t6+i569OjKsmUf4uHhwZkzZ2jWrB27d2fwyiuzeeedN67uUYiIyEXGjfs/3nrrHSwWi+2cNPYWfKFvk0xOTiExcTUAM2Y8j4eHBwAeHh5ERLQDIDNzX7HDi4jI5X366TKGDLkbgFdfnUqXLp14+ulH7RpbaMFfeJN99erVCA1tUeC2Q4cOA1CnToBtW1xcPBaLH4MG3XPZOePi4gkPjyI8PIpjx07YFVRE5Hr0xx/ZdOp0EwABAf7cddftxMXNt2tsoYdojhw5CkDVqlUKbM/JySE5Of/9pJGREQD8+eefTJnyEp6enlecMzZ2JLGxIwFo2jQU+MOusCIi1xt//5qcPXsWf/9ajB79T06fPoOvbyW7xha6gr/w5vo9e/aSlfXXoZjJk6dz/PhxAgL8GTZsIJC/Mg8La0lAQK3iPA4REfkfzz//FMHBQbz88hS8vb2pXNmXmTNfsGtsoQU/eHB/GjSox5kzZ4iM7EVs7EN06tSb6dNn4evry7Jli/Dx8SEvL4+pU2cyefKTV/2AREQEzp07xw8//IinpwcDB97JwYPpHDiwk0GD+ts1vtCCr1SpEqtXLycmZgB5eXnMnTufpKR1dO3ambS0DbRpEwbA3LnziYxsb3tPvDGG8+fPX8VDExG5vrm7u/PZZ1/wyy8ZxRpv19skAwPrsGBBHACbNm0hMrInqak7cXd3t+2ze3cGS5Z8xpIlnwGwd28WjRs34vnnnypWMBERgaiojkyaNI0///yzwOHvO+/sW+jYIp+qoG3bMBYvnsfmzVvZsmUbPXp0BeCRRx4gJib/WHzfvoNp2bIF999/5XMVi4jIlc2b9z4ADz74OJB/dMRisXDu3NFCxxbrXDR9+95K3763FthWt24gdesGArB/f1pxphURkf8xceJjWCyWYo3VycZERMqwZ5+dAEB2djbly5cv9G3of6cP/BARKcMyMvbSrl0XqldvxPffr+Xmm29l4sTCT1MAKngRkTJtzJiH+fXXAxhjcHOz0LlzBxYt+sSusSp4EZEybO3aDTzwwL22640aNWTfvv12jVXBi4iUYdWrV2X79h0A/P77YRYt+oTatf3tGqtfsoqIlGH33juCJ5+cDMDQofkr+alTn7VrrApeRKQULFz4EUOGjAbgoYfGMHPmlT9e73ImTHiY2rX9+eKLlQDcdltPhg8fbNdYFbyISAnbt+9X7r//EcqVK8fZs2evaq7x4/9NTMxARowYUuSxOgYvIlKCjDGMGDGW2rUD6N+/8NMJFGbWrDm0b9+N5s1vZMqUl8jI2Gv3WBW8iEgJmjlzNklJ63n//Ti8vb2uer7t29cxceJjeHiU4+mnp9Co0Q106tTbrrEqeBERO61dm2z7NLrw8Cji4uIL3L59eyoTJkxi0qQnCAtrVSL32aJFM5555t+sWPEx48aNtuWwh47Bi4jYqUOH9ixaNO+yt3/88eecPn2aVavWsHr1OrZu3Q7A558vx8fHhxdeeKbI9zlz5mw++uhz1q/fyPnz52nUqCFDhtxl11gVvIhICTHGYIxh+fKvC2zfs2cv69ZtKNacDz/8JDVqVGfs2HuIiRlI+/bhdo/VIRoRkRLy7LMTMCbbdhkxIv/tjA89NAar9YtizfnFF4vZvz+N116bTvv24Wzfnmr3uWgctoJPSFhOQsIKcnLOYLW+5KgYxZKbm4vVanV0DLvl5p7FmnTQ0TGKxNkyO1tegNwT553qdXy96t37FtLS0vnww09YvPgz0tLSAZg0qfCPR3VYwUdH9yY6ujd164bSpcv/OSpGsSQmJhIVFeXoGHazJn5JVEf7/rS5rLAmHXSqzNakg9Q5MM7RMYrk14A3iIr0dXQMlxYf/ybx8W8Wa+zPP//C4sWf8uGHn/DTT2m2D/ro06cHw4YNsmsOHYMXESmDmjYNx2KxEBDgz7hxo7nxxrYMHz6G0aOHX/SBS5ejghcRKaPc3Ny4+eZIunbtTNOmjYs+vhQyiYjIVXrttWl06HAjH374Cf37D6dNm5uxWCxs3LiZI0cK/zxWUMGLiJRJ48bdy6pVX5KV9RMvvzyFNm1aAzBlygz8/ZvYNYcKXkSkDAsI8Oehh8ayZs1X7N27jZdemkzbtmF2jVXBi4g4icDAOjz88AOsX/+NXfur4EVEXJQKXkTERangRURclApeRMRFqeBFRFyUCl5ExEWp4EVEXJTORSMi142Qo9v59IOGxRr7ybpcaOFcZ77VCl5ExEWp4EVEXJQKXkTERdld8Dk5OUycOIWQkAjKlw8gIKAp/foNITU1rTTziYhIMdlV8Pv3HyA8vAuTJ0/HGMOwYQOpUaMaS5d+Sbt2Xdm27afSzikiIkVkV8HHxMSSnr6LHj26snVrEnPmzGTTplUEBTXg5MmTvPLK7NLOKSIiRVRowScnp5CYuBqAGTOex8PDAwAPDw8iItoBkJm5rxQjiohIcRRa8FZrEgDVq1cjNLRFgdsOHToMQJ06AWRk7MVi8bNdQkNvuuyccXHxhIdHER4exYkTJ64mv4iIXEahf+h04bP/qlatUmB7Tk4OycmbAIiMjLBtX7fuawIDa9tW+pcSGzuS2NiRANStG8off/xR5OAiInJlha7gg4ODANizZy9ZWX8dipk8eTrHjx8nIMCfYcMG2rb36TOA7t37sWrVmlKIKyIi9iq04AcP7k+DBvU4c+YMkZG9iI19iE6dejN9+ix8fX1ZtmwRPj4++Pr68v77c0lMTKBJk0aMGDHW7k/+FhGRkldowVeqVInVq5cTEzOAvLw85s6dT1LSOrp27Uxa2gbatMn/8NeqVaswZMjdtGoVyqhRQzl16hS7d2eUdn4REbkMu042FhhYhwUL4gDYtGkLkZE9SU3dibu7u22flSu/Y9++X+nQoT0ffPAR3t7eBAU1KJXQIiJSuCKfTbJt2zAWL57H5s1b2bJlGz16dAWgYsUKTJs2i4yMTAIDa7NgwRyqVata4oFFRMQ+xTpdcN++t9K3760FtnXo0J60tI0lEkpERK6eTjYmIuKiVPAiIi5KBS8i4qJU8CIiLkoFLyLiolTwIiIuSgUvIuKiVPAiIi5KBS8i4qJU8CIiLkoFLyLiolTwIiIuSgUvIuKiVPAiIi6qWKcLLgkJCctJSFjBuXPnSExMdFSMYsnNzcVqtTo6ht1yT5zHuua4o2MUibNlzj1xnl8D3nB0jCLJzT2LNemgo2NIKXJYwUdH9yY6ujdNm95IVFSUo2IUi9VqdarMzpYXnC+z1WolKtLX0TGKxJp0kKiO/o6O4VJGj/4na9Ykk5X1K15enrRvH860ac8RGtrCIXl0iEZEpIT8978L8POrzODB/fH1rcTy5V/Tq9ddnDp1yiF5HLaCFxFxNSkpVtq2DQMgI2MvDRu25tdf95OamkabNmHXPI9W8CIidlq7Npnw8CjbJS4uvsDtF8od4PTpMwC4u7sTEOCYQ2FawYuI2KlDh/YsWjSv0P1yc3MZNWocAA8/PM5hBa8VvIhICTp06DBdukSzdm0y9947ghdffM5hWVTwIiIlZO/eTDp27EVKyg9MmPAwcXGvYrFYHJZHh2hEREpIhw492b//APXqBXLy5EnGj/83AEOG3M2NN7a95nlU8CIiJWT//gMAZGbu49VX37JtDwtrqYIXEXFmxmQ7OkIBOgYvIuKiVPAiIi5KBS8i4qJ0DF5ErhteVYJofPf7xRrrn5fAsRMlHKiUaQUvIuKiVPAiIi5KBS8i4qJU8CIiLsrugs/JyWHixCmEhERQvnwAAQFN6ddvCKmpaaWZT0REismugt+//wDh4V2YPHk6xhiGDRtIjRrVWLr0S9q168q2bT+Vdk4RESkiuwo+JiaW9PRd9OjRla1bk5gzZyabNq0iKKgBJ0+e5JVXZpd2ThERKaJCCz45OYXExNUAzJjxPB4eHgB4eHgQEdEOyD+xTlHExcXbPhHl2LFjRc0sIiJ2KLTgrdYkAKpXr3bRJ4MfOnQYgDp1AgBYvXotYWEd8fHxJyys42XnjI0dSUqKlZQUK5UrVy52eBERubxCC/7IkaMAVK1apcD2nJwckpM3ARAZGUF2djbR0YMICWlOSkoio0cPL4W4IiJir0JPVRAcHATAnj17ycraR926gQBMnjyd48ePExDgz7BhA/nkkwSOHTvO1KnPULduICEhzUs3uYiIXFGhK/jBg/vToEE9zpw5Q2RkL2JjH6JTp95Mnz4LX19fli1bhI+PD1lZv2KxWBgwYBQBAU156KHHMcZci8cgIiKXUGjBV6pUidWrlxMTM4C8vDzmzp1PUtI6unbtTFraBtq0CQPyD+EYYxg06E7Gjx/LrFlzWLnyu1J/ACIicml2vU0yMLAOCxbEcejQL6SkWPHy8iI1dSfu7u62fbp1uxk3Nzc8PT3x8vIEwNPTo3RSi4hIoYp8uuC2bcNYvHgemzdvZcuWbfTo0RWARo0aMnv2DCZNmsapU6d49NEH6dKlc4kHFhER+xTrfPB9+95K3763XrT9vvtGcd99o646lIiIXD2dbExExEWp4EVEXJQKXkTERangRURclApeRMRFqeBFRFyUCl5ExEWp4EVEXJQKXkTERangRURclApeRMRFqeBFRFyUCl5ExEWp4EVEXFSxThdcEhISlpOQsIKsrEzCw28o8fkPHTpCjRrVSnxegGPHjlG5cuUSn7e0MpdWXnC+zM6WF5wvc2l+76Wn7yqVeV2Vwwo+Oro30dG9AYiLe7XE5w8PjyIlxVri8wLExj7kVJlLKy84X2ZnywvOl7m0v/fEfg4/RBMd3cvREYrM2TI7W15wvszOlheU+XpQBgq+t6MjFJmzZXa2vOB8mZ0tLyjz9cDhBV9aYmNHOjpCkSlz6XO2vOB8mZ0tryuzGJNtHB1CRORaaNsmgk3J7xdr7Px3Ezh2wpsHHxxTwqlKj8uu4EVErncqeBERF6WCFxFxUSp4EREXpYIXESlBp06d4p//fJSaNYPx8fEnMrInyckpDsmighcRKUHjx0/g9dfnUqtWTfr168O6dRu45ZY7OHz4yDXPooIXESkhv/9+iHfeeQ83Nze+/XYpCxf+l6FD7yYnJ4fXX4+75nlU8CJy3fD29mL//kPFGvvj9p9JTd1JeHiU7RIXF19gn59+2sGZM2eoVy+QmjVrANhOprhly7aryl4cDjvZmIjItfbwQ0N58aV4Xn350SKN+/33o1hXbWLDxjW4u7tfdr/ffsv/4VGxYkXbtgoVKgBw8ODvxUh8dbSCF5Hrxh13j2Z98rYir+Knv/wuj4yPuWK5A9Sqlb9qz83NtW278LW/f80ipr16KngRuW64ubnx2CMjePGleLvH/P77Ub5L3MjAoWML3bdFi2Z4eHiQmbmP337LX7Fv3PgDAK1bhxYr89VQwYvIdaWoq3h7V+8AtWrVZOTIIZw/f55u3W5n0KB7WLjwIypWrMgDD8RebfQiU8GLyHWlKKv4oqzeL3j11ancf/9ofvvtdz777AsiItqxcuUn1KhR/SpSF4/OJiki153z589zU0QnPl0yg9q1a1x2v0f/PZMbWjdlyPAHrmG6kuMSK/icnBwmTpxCSEgE5csHEBDQlH79hpCamuboaJfkbHnB+TI7W15Q5mvJnlV8cVbvZY3Tr+D37z9Aly7RpKfvonnzpnTqdBPr1m1g27ZUypcvz/r1X9OyZYijY9o4W15wvszOlheU2REKW8U7++odXGAFHxMTS3r6Lnr06MrWrUnMmTOTTZtWERTUgJMnT/LKK7MdHbEAZ8sLzpfZ2fKCMjvClVbxrrB6Bycv+OTkFBITVwMwY8bzeHh4AODh4UFERDsAMjP3OSzf/3K2vOB8mZ0tLyizI13uHTVFeedMWebUBW+1JgFQvXo1QkNbFLjt0KHDANSpE8CZM2eIiOiOh0d1/P2bXPOcF9ib12pdTdOm4Xh716Jx4zZ89dW31zzrBfZmTk1No0GDlnh51SQoqDUff7z0mmcF+/NeEBcXj8Xix6BB91y7kP/D3swZGXuxWPxsl9DQm6551guK8jyvXr2WsLCO+Pj4ExbW8doGLcSlVvGusnoHJy/4I0eOAlC1apUC23NyckhO3gRAZGQEFouFO+64jZtvjrzmGf/O3rynTv3JhAn/YvPmVVSu7Mt9942/5lkvsDezr28l3n77NbZsWU3FihV48snnr3lWsD8vwJ9//smUKS/h6el5bUP+j6JkBli37muysn7i228d80MU7M+cnZ1NdPQgQkKak5KSyOjRw6951sL87yreVVbv4OQFHxwcBMCePXvJyvrrn4OTJ0/n+PHjBAT4M2zYQMqVK8fjj48nMLC2o6IC9uft1as7I0cOpUWLZrRtG8bRo9mOimx35sDAOnTvHkVwcBC+vr40b+6YfynZmxfyV+9hYS0JCKjlkKwXFCUzQJ8+A+jevR+rVq255lkvsDfzF1+s5Nix40yd+gwhIc0d8sc+hfn7Kt6VVu/g5O+iycnJoVWrSDIyMqlbN5BevbqxY0c6SUnr8PX1JTHxc9q0CbPtP3LkWFas+JaDB9OdIu/27alERNzCmDGjeOklx6yIi5J59uy3GT9+AgDz5r3B0KEDymzevLw8goPbsHz5Evr2HUxERDsWLXrnmuctSuajR/9gxYpvCA1tzlNPPc/XX1vZty+VatWqltnMU6e+whNPTKJ9+3AyMjIZMKAfM2dOxWKxXPPMV3LhHTWNg+txa69Ip37nzN859Qq+UqVKrF69nJiYAeTl5TF37nySktbRtWtn0tI2FCjLsqAoeX/++RduueUOOnW6ialTn3WKzEOH3k1KSiIdOtzI2LGPYMy1XzvYm3fu3PlERranRYtmABhjOH/+/DXPW5TMVatWYciQu2nVKpRRo4Zy6tQpdu/OKPOZjTEMGnQn48ePZdasOaxc+Z1DMl/JhVX8j9t+dpnVOzh5wQMEBtZhwYI4Dh36hZQUK15eXqSm7rzo+FlaWjrHjh3n3LlzpKWlc/LkyTKbNytrH927307NmtWZNetFDh78jXPnzjkkr72Z16xZz44d6Xh7e+Pp6Ym3t5fDVmn25N29O4MlSz7Dw6M6e/dmsXjxp0yc+B+H5LU388qV3/HOOwtIS0vngw8+wtvbm6CgBmU6c7duN+Pm5oanpydeXvm/6/D09HBU5CvqPzCWH7eluMSx9wuc+hDNpXz++Zds3ryVDh3a06NHV9t2i8WvwH6JiQlERXW61vEucqm88fHvM2rUuAL77dmzlQYN6jsi4kUulXnBgkU88shTHD+eQ3BwENOnT6J371scnDTfpfJmZe2znbu7b9/BtGzZgnnz3qB27YArTXXNXCrz2rXJ3HPPA2RkZBIYWJupU5/lrrtud3DSv1zue2/OnHlMmjSNU6dO8Y9/DGPatEkOTHl9cbmCFxGRfE5/iEZERC5NBS8i4qJU8CIiLkoFLyLiolTwIiIuSgUvIuKiVPAiIi5KBS8i4qL+H/Eo1MVf3yVpAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x432 with 2 Axes>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 297
        },
        "id": "8tvCA6D-21U6",
        "outputId": "4500daf0-411a-4f89-c843-2f34558b70ce"
      },
      "source": [
        "!pip install hickle\n",
        "import hickle as hk\n",
        "from google.colab import files\n",
        "hk.dump(FinalTimeSeries.cpu().detach().numpy(),'TS_FM_2.hkl', mode='w')\n",
        "hk.dump(FinalConstants.cpu().detach().numpy(),'C_FM_2.hkl', mode='w')\n",
        "hk.dump(InteractionsFinal.cpu().detach().numpy(),'I_FM_2.hkl', mode='w')\n",
        "files.download('TS_FM_2.hkl')\n",
        "files.download('C_FM_2.hkl')\n",
        "files.download('I_FM_2.hkl')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Collecting hickle\n",
            "  Downloading hickle-4.0.4-py3-none-any.whl (49 kB)\n",
            "\u001b[?25l\r\u001b[K     |██████▊                         | 10 kB 26.9 MB/s eta 0:00:01\r\u001b[K     |█████████████▍                  | 20 kB 20.1 MB/s eta 0:00:01\r\u001b[K     |████████████████████            | 30 kB 11.2 MB/s eta 0:00:01\r\u001b[K     |██████████████████████████▊     | 40 kB 9.1 MB/s eta 0:00:01\r\u001b[K     |████████████████████████████████| 49 kB 2.4 MB/s \n",
            "\u001b[?25hRequirement already satisfied: numpy>=1.8 in /usr/local/lib/python3.7/dist-packages (from hickle) (1.19.5)\n",
            "Collecting h5py<3.0.0,>=2.8.0\n",
            "  Downloading h5py-2.10.0-cp37-cp37m-manylinux1_x86_64.whl (2.9 MB)\n",
            "\u001b[K     |████████████████████████████████| 2.9 MB 6.1 MB/s \n",
            "\u001b[?25hRequirement already satisfied: dill>=0.3.0 in /usr/local/lib/python3.7/dist-packages (from hickle) (0.3.4)\n",
            "Requirement already satisfied: six>=1.11.0 in /usr/local/lib/python3.7/dist-packages (from hickle) (1.15.0)\n",
            "Installing collected packages: h5py, hickle\n",
            "  Attempting uninstall: h5py\n",
            "    Found existing installation: h5py 3.1.0\n",
            "    Uninstalling h5py-3.1.0:\n",
            "      Successfully uninstalled h5py-3.1.0\n",
            "Successfully installed h5py-2.10.0 hickle-4.0.4\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "\n",
              "    async function download(id, filename, size) {\n",
              "      if (!google.colab.kernel.accessAllowed) {\n",
              "        return;\n",
              "      }\n",
              "      const div = document.createElement('div');\n",
              "      const label = document.createElement('label');\n",
              "      label.textContent = `Downloading \"${filename}\": `;\n",
              "      div.appendChild(label);\n",
              "      const progress = document.createElement('progress');\n",
              "      progress.max = size;\n",
              "      div.appendChild(progress);\n",
              "      document.body.appendChild(div);\n",
              "\n",
              "      const buffers = [];\n",
              "      let downloaded = 0;\n",
              "\n",
              "      const channel = await google.colab.kernel.comms.open(id);\n",
              "      // Send a message to notify the kernel that we're ready.\n",
              "      channel.send({})\n",
              "\n",
              "      for await (const message of channel.messages) {\n",
              "        // Send a message to notify the kernel that we're ready.\n",
              "        channel.send({})\n",
              "        if (message.buffers) {\n",
              "          for (const buffer of message.buffers) {\n",
              "            buffers.push(buffer);\n",
              "            downloaded += buffer.byteLength;\n",
              "            progress.value = downloaded;\n",
              "          }\n",
              "        }\n",
              "      }\n",
              "      const blob = new Blob(buffers, {type: 'application/binary'});\n",
              "      const a = document.createElement('a');\n",
              "      a.href = window.URL.createObjectURL(blob);\n",
              "      a.download = filename;\n",
              "      div.appendChild(a);\n",
              "      a.click();\n",
              "      div.remove();\n",
              "    }\n",
              "  "
            ],
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "download(\"download_69810306-70de-455f-a4ba-2d0416b47dad\", \"TS_FM_2.hkl\", 780164)"
            ],
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "\n",
              "    async function download(id, filename, size) {\n",
              "      if (!google.colab.kernel.accessAllowed) {\n",
              "        return;\n",
              "      }\n",
              "      const div = document.createElement('div');\n",
              "      const label = document.createElement('label');\n",
              "      label.textContent = `Downloading \"${filename}\": `;\n",
              "      div.appendChild(label);\n",
              "      const progress = document.createElement('progress');\n",
              "      progress.max = size;\n",
              "      div.appendChild(progress);\n",
              "      document.body.appendChild(div);\n",
              "\n",
              "      const buffers = [];\n",
              "      let downloaded = 0;\n",
              "\n",
              "      const channel = await google.colab.kernel.comms.open(id);\n",
              "      // Send a message to notify the kernel that we're ready.\n",
              "      channel.send({})\n",
              "\n",
              "      for await (const message of channel.messages) {\n",
              "        // Send a message to notify the kernel that we're ready.\n",
              "        channel.send({})\n",
              "        if (message.buffers) {\n",
              "          for (const buffer of message.buffers) {\n",
              "            buffers.push(buffer);\n",
              "            downloaded += buffer.byteLength;\n",
              "            progress.value = downloaded;\n",
              "          }\n",
              "        }\n",
              "      }\n",
              "      const blob = new Blob(buffers, {type: 'application/binary'});\n",
              "      const a = document.createElement('a');\n",
              "      a.href = window.URL.createObjectURL(blob);\n",
              "      a.download = filename;\n",
              "      div.appendChild(a);\n",
              "      a.click();\n",
              "      div.remove();\n",
              "    }\n",
              "  "
            ],
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "download(\"download_be45529f-5cfe-4f4f-a060-491b4de6396d\", \"C_FM_2.hkl\", 6224)"
            ],
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "\n",
              "    async function download(id, filename, size) {\n",
              "      if (!google.colab.kernel.accessAllowed) {\n",
              "        return;\n",
              "      }\n",
              "      const div = document.createElement('div');\n",
              "      const label = document.createElement('label');\n",
              "      label.textContent = `Downloading \"${filename}\": `;\n",
              "      div.appendChild(label);\n",
              "      const progress = document.createElement('progress');\n",
              "      progress.max = size;\n",
              "      div.appendChild(progress);\n",
              "      document.body.appendChild(div);\n",
              "\n",
              "      const buffers = [];\n",
              "      let downloaded = 0;\n",
              "\n",
              "      const channel = await google.colab.kernel.comms.open(id);\n",
              "      // Send a message to notify the kernel that we're ready.\n",
              "      channel.send({})\n",
              "\n",
              "      for await (const message of channel.messages) {\n",
              "        // Send a message to notify the kernel that we're ready.\n",
              "        channel.send({})\n",
              "        if (message.buffers) {\n",
              "          for (const buffer of message.buffers) {\n",
              "            buffers.push(buffer);\n",
              "            downloaded += buffer.byteLength;\n",
              "            progress.value = downloaded;\n",
              "          }\n",
              "        }\n",
              "      }\n",
              "      const blob = new Blob(buffers, {type: 'application/binary'});\n",
              "      const a = document.createElement('a');\n",
              "      a.href = window.URL.createObjectURL(blob);\n",
              "      a.download = filename;\n",
              "      div.appendChild(a);\n",
              "      a.click();\n",
              "      div.remove();\n",
              "    }\n",
              "  "
            ],
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "download(\"download_d4dc2a25-4918-489f-aa04-bccf82b254f1\", \"I_FM_2.hkl\", 6432)"
            ],
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ]
          },
          "metadata": {}
        }
      ]
    }
  ]
}