ForMIC: Foraging via Multiagent RL with Implicit Communication

Key Files

• DRLMAF_A3C_RNN.py: Multi-agent training code
• ACNet.py: defines network architecture
• near_optimal_driver.py: baseline centralized near optimal model
• cardinality_policy.py: baseline decentralized model
• csaf_policy.py: baseline decentralized model
• csaf_policy_11.py: baseline decentralized model taking advantage of 11x11 FOV

Other Links

• fully trained model
• scenarios used for testing

Authors

Samuel Shaw

Emerson Wenzel

Lexi Walker

Guillaume Sartoretti