2D simulation of a configurable viral pandemic developed under Unity 2019 ECS DOTS Environment.
Each entity replicates the human behaviour, characterized by some primary basic needs such as going to work, at home to rest, to the grocery shop but also social needs like need for a run in a park or visiting friends.
The purpose of the project is to simulate the trend of a viral infection enhancing the parallelism provided by ECS: many entities are spawn at the beginning, the majority is non-infected. The intention is to show how a potential viral infection can spread among the population of entities according both to viral parameters but also on entities behaviour.
-
Humans are represented with a 2d sprite symbol.
-
The color of the symbol changes with the “infection state” in which the human is.
- The simulation follows the SEIRS model, it means that the contagion follow a state machine of the kind:
- S susceptible
- E exposed
- I infected, modeled as Symptomatic and Asymptomatic
- R recovered or removed, the entity may also die
- S susceptible, after some time after recovery, an entity may be infected again
- Humans movement based on pathfinding in order to satisfy needs, each time a human needs to satisfy a demand, the pathfinding algorithm drives the human to the closest place where to satisfy the necessity:
| _ | Home | Park | Pub* | Supermarket | Office |
|---|---|---|---|---|---|
| Fatigue | X | ||||
| Hunger | X | X | |||
| Sportivity | X | ||||
| Sociality | X* | X | X | ||
| Grocery | X | ||||
| Work | X |
The Home in the Sociality need is not the Human home, in order to simulate visit to friends. If the lockdown setting is True, all needs increase more slowly and are correlated to the human social responsibility. Furthermore, the Pubs are closed.
Needs are set in order to reproduce a normal behavior:
- Fatigue: Human rest for 8 hours, need restored after 16 hours from fulfillment;
- Hunger: Human eat for 1 hour three times each day
- Sportivity: Human does 1.5 hours of sport every two days
- Sociality: Human does 2 hours of sociality each day
- Grocery: Human goes to the supermarket for one hour once every 3 days
- Work: Human goes to work for 8 hourse every day
The map is implemented using a tilemap file as input, specified in the configuration file. Each sprite model a specific city element:
Contagion probability based formula that keeps into account several parameters such as:
- Distance, 2 meters is the value of risk
- Time, 15 minutes are enough to make infection happen
- Social Responsibility (that represents several factors about social behaviours during a pandemic, mainly mask usage and social distancing)
./Conf/conf.json is the configuration file, the user can set up many parameters for modeling the simulation, such as
- number of humans i.e. population
- number of initial infects
- time scale i.e. speed-up time factor
- probability of death
- length (among min and max value) of every state of infection
- map to use (see folder ./Conf/Maps)
- lockdown, set to true to see an implementation of a partial lockdown
For avoiding main memory bottleneck, we suggest to execute the simulation with at least 16 GB DRAM and the following parameters less than a cap value:
- numberOfHumans at most 250k-300k
- timeScale at most 60
- Upper right corner shows stats
- Key "Q" -> Zoom In
- Key "E" -> Zoom Out
- W-A-S-D for xy movement on the map
No restriction simulation example:
Lockdown simulation example:
