-
Notifications
You must be signed in to change notification settings - Fork 99
BatchSpecific
Julien Mazars edited this page Jan 15, 2016
·
19 revisions
When an experiment of type Batch is run, a dedicated UI is displayed, depending on the parameters to explore and of the exploration methods.
In batch mode, the top information bar displays 3 distinct information (instead of only the cycle number in the GUI experiment):
-
Run: the run number. One run corresponds to X executions of simulation with one given parameters values (X is an integer given by the facet
repeat
in the definition of the exploration method); -
Simulation: the number of replications done (and the number of replications specified with the
repeat
facet); - Cycle: the cycle number in the current simulation.
The parameters view is also a bit different in the case of a Batch UI. The following interface is generated given the following model part:
experiment Batch type: batch repeat: 2 keep_seed: true until: (food_gathered = food_placed) or (time > 400) {
parameter 'Size of the grid:' var: gridsize init: 75 unit: 'width and height';
parameter 'Number:' var: ants_number init: 200 unit: 'ants';
parameter 'Evaporation:' var: evaporation_rate among: [0.1, 0.2, 0.5, 0.8, 1.0] unit: 'rate every cycle (1.0 means 100%)';
parameter 'Diffusion:' var: diffusion_rate min: 0.1 max: 1.0 unit: 'rate every cycle (1.0 means 100%)' step: 0.3;
method exhaustive maximize: food_gathered;
The interface summarizes all model parameters and the parameters given to the exploration method:
- Environment and Population: displays all the model parameters that should not be explored;
-
Parameters to explore: the parameters to explore are the parameters defined in the experiment with a range of values (with
among
facet ormin
,max
andstep
facets); - Exploration method: it summarizes the Exploration method and the stop condition. For exhaustive method it also evaluates the parameter space. For other methods, it also displays the method parameters (e.g. mutation or crossover probability...). Finally the best fitness found and the last fitness found are displayed (with the associated parameter set).
- Installation and Launching
- Workspace, Projects and Models
- Editing Models
- Running Experiments
- Running Headless
- Preferences
- Troubleshooting
- Introduction
- Manipulate basic Species
- Global Species
- Defining Advanced Species
- Defining GUI Experiment
- Exploring Models
- Optimizing Model Section
- Multi-Paradigm Modeling
- Manipulate OSM Data
- Diffusion
- Using Database
- Using FIPA ACL
- Using BDI with BEN
- Using Driving Skill
- Manipulate dates
- Manipulate lights
- Using comodel
- Save and restore Simulations
- Using network
- Headless mode
- Using Headless
- Writing Unit Tests
- Ensure model's reproducibility
- Going further with extensions
- Built-in Species
- Built-in Skills
- Built-in Architecture
- Statements
- Data Type
- File Type
- Expressions
- Exhaustive list of GAMA Keywords
- Installing the GIT version
- Developing Extensions
- Introduction to GAMA Java API
- Using GAMA flags
- Creating a release of GAMA
- Documentation generation