This repository implements a PandemicSimulator agent-based pandemic simulator. The main motivation behind the simulator is to model the effects of sequential decisions that can be made to manage the impact of COVID-19 virus and identify effective re-opening strategies. The simulator comprises of multiple locations with different characteristics, different categories of people, and their realistic stateful behaviors that differ for each category. An OpenAI Gym environment is also made available to optimize a reopening strategy based on Reinforcement Learning.
See https://docs.google.com/document/d/16_LwPFMKI1Hqts6GeFZvqSJ2zKNzDWpemWUb8Oyrn_k for information on parameters used in the model.
/docs -- API documentation plus hierarchy
/python/pandemic_simulator/ -- Source files
/scripts -- Example scripts to run the simulator and the OpenAI gym environment
To run the simulator, just execute the following in a bash (or equivalent) shell:
python3 scripts/run_pandemic_sim.py
The code runs the simulator under stage-0 policy (see http://austintexas.gov/covid19 for more info on stages).
Similarly, to run the OpenAI Gym env, execute the following:
python3 scripts/run_pandemic_env.py
If you have any questions regarding the code, please feel free to contact Roberto Capobianco ([email protected]) or Varun Kompella ([email protected]).