This paper proposes a definition of system health in the context of multiple agents optimizing a joint reward function. We use this definition as a credit assignment term in a policy gradient algorithm to distinguish the contributions of individual agents to the global reward. The health-informed credit assignment is then extended to a multi-agent variant of the proximal policy optimization algorithm and demonstrated on particle and multiwalker robot environments that have characteristics such as system health, risk-taking, semi-expendable agents, continuous action spaces, and partial observability. We show significant improvement in learning performance compared to policy gradient methods that do not perform multi-agent credit assignment.
@article{arxiv.1908.01022,
title = {Health-Informed Policy Gradients for Multi-Agent Reinforcement Learning},
author = {Ross E. Allen and Jayesh K. Gupta and Jaime Pena and Yutai Zhou and Javona White Bear and Mykel J. Kochenderfer},
journal= {arXiv preprint arXiv:1908.01022},
year = {2021}
}