Distributed Value Decomposition Networks with Networked Agents

TL;DR

This paper introduces distributed value decomposition networks (DVDN) for multi-agent reinforcement learning, enabling decentralized training with communication among agents, and demonstrates their effectiveness across multiple tasks and environments.

Contribution

The paper proposes DVDN and DVDN (GT), novel algorithms for decentralized multi-agent RL that do not require centralized training, addressing communication constraints.

Findings

DVDN approximates centralized value decomposition performance.

Both algorithms perform well despite communication-induced information loss.

Validated across ten tasks in three environments.

Abstract

We investigate the problem of distributed training under partial observability, whereby cooperative multi-agent reinforcement learning agents (MARL) maximize the expected cumulative joint reward. We propose distributed value decomposition networks (DVDN) that generate a joint Q-function that factorizes into agent-wise Q-functions. Whereas the original value decomposition networks rely on centralized training, our approach is suitable for domains where centralized training is not possible and agents must learn by interacting with the physical environment in a decentralized manner while communicating with their peers. DVDN overcomes the need for centralized training by locally estimating the shared objective. We contribute with two innovative algorithms, DVDN and DVDN (GT), for the heterogeneous and homogeneous agents settings respectively. Empirically, both algorithms approximate the performance of value decomposition networks, in spite of the information loss during communication, as demonstrated in ten MARL tasks in three standard environments.