Corpus ID: 199442485

Health-Informed Policy Gradients for Multi-Agent Reinforcement Learning

@article{Allen2019HealthInformedPG,
  title={Health-Informed Policy Gradients for Multi-Agent Reinforcement Learning},
  author={R. Allen and J. W. Bear and Jayesh K. Gupta and Mykel J. Kochenderfer},
  journal={ArXiv},
  year={2019},
  volume={abs/1908.01022}
}
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 simple particle environments that have characteristics such as system… Expand
Toward an Autonomous Aerial Survey and Planning System for Humanitarian Aid and Disaster Response
Deep Implicit Coordination Graphs for Multi-agent Reinforcement Learning

References

SHOWING 1-10 OF 32 REFERENCES
Multi-Agent Actor-Critic for Mixed Cooperative-Competitive Environments
Cooperative Multi-agent Control Using Deep Reinforcement Learning
Counterfactual Multi-Agent Policy Gradients
Proximal Policy Optimization Algorithms
Credit Assignment For Collective Multiagent RL With Global Rewards
Learning sequences of actions in collectives of autonomous agents
Policy Gradient Methods for Reinforcement Learning with Function Approximation
Trust Region Policy Optimization
An Optimal Best-First Search Algorithm for Solving Infinite Horizon DEC-POMDPs
High-Dimensional Continuous Control Using Generalized Advantage Estimation
...
1
2
3
4
...