Distributed On-Policy Actor-Critic Reinforcement Learning
Abstract
In this paper, a novel distributed on-policy Actor-Critic algorithm for multiagent reinforcement learning is proposed.The algorithm consists of the temporal difference scheme with function approximation at the Critic stage, and a policy gradient algorithm at the Actor stage, derived starting from a global objective.At both stages, decentralized agreement among the agents is achieved using the linear dynamic consensus strategy.Compared to the existing schemes, the algorithm has improved convergence rate and noise immunity, and a possibility to achieve multi-task global optimization.
Metrics
Topics
Related Documents
A Multi-Agent Off-Policy Actor-Critic Algorithm for Distributed Reinforcement Learning
Wesley A. SuttleZhuoran YangKaiqing ZhangZhaoran WangTamer BaşarJi Liu
Distributed Multi-Agent Reinforcement Learning by Actor-Critic Method
Multi-agent Gradient-Based Off-Policy Actor-Critic Algorithm for Distributed Reinforcement Learning
Multi-agent off-policy actor-critic algorithm for distributed multi-task reinforcement learning
Miloš S. StankovićMarko BekoNemanja IlićSrdjan Stanković
Supervised Actor-Critic Reinforcement Learning
Michael T. RosensteinAndrew G. BartoJennie SiAndy BartoWarren B. PowellDonald C. Wunsch