This project simulates a multi-agent system with agents navigating and coordinating on a graph. The codebase is primarily written in Python, with supporting scripts in JavaScript and HTML.
The graph considered is an anonymous port labeled graph. The agents can move from one node to another in synchrnous rounds.
WIP: Near-Linear Time Dispersion by Sudo et al.
agent.py: Defines theAgentclass and related enums (as dictionaries) for agent roles, phases, statuses, and node statuses. Contains the agent logic and state transitions.simulation_wrapper.py: Runs the main simulation loop, initializing agents and stepping through rounds.main.py: (If present) Likely handles overall orchestration or entry point for running the simulation.simulation-runner.js: JavaScript file for running or visualizing simulations (details depend on implementation).index.html: Web interface for running or visualizing the simulation (if applicable).
- Agents: Each agent has a state including
status,role,phase, and location. State changes are logged for traceability. - Graph: Agents move and interact on a graph, with nodes and ports representing locations and connections.
- Simulation Loop: The simulation proceeds in rounds, with each agent computing its actions and updating its state.
Run the simulation directly in the browser. The project utilizes Pyodide to run python compiled into webassembly for the browser. The website is at drdebmath.github.io/CCMModel
- Modify
agent.pyto change agent behavior, roles, or state transitions. - Update
simulation_wrapper.pyto change the simulation parameters or loop. - Use or adapt
index.htmlandsimulation-runner.jsfor visualization or web-based interaction.
- Agent state changes are printed to the console for debugging and traceability.
- Use the print statements to follow the simulation's progress and agent decisions.
MIT
For questions or contributions, contact @drdebmath.