TechRxiv
Decentralized_Multi_Agent_Advantage_Actor_Critic_Abstract.pdf (251.67 kB)
Download file

Decentralized Multi-Agent Advantage Actor-Critic

Download (251.67 kB)
preprint
posted on 22.02.2022, 00:56 authored by Scott BarnesScott Barnes
We present a decentralized advantage actor-critic algorithm that utilizes learning agents in parallel environments with synchronous gradient descent. This approach decorrelates agents’ experiences, stabilizing observations and eliminating the need for a replay buffer, requires no knowledge of the other agents’ internal state during training or execution, and runs on a single multi-core CPU.

History

Email Address of Submitting Author

scottgbarnes@gwu.edu

Submitting Author's Institution

The George Washington University

Submitting Author's Country

United States of America

Usage metrics

Licence

Exports