CM3: Cooperative Multi-goal Multi-stage Multi-agent Reinforcement Learning

TL;DR

CM3 introduces a novel two-stage curriculum and credit assignment mechanism to improve cooperative multi-goal multi-agent reinforcement learning, enabling faster learning in complex multi-agent tasks.

Contribution

The paper proposes a new multi-goal multi-agent policy gradient with a credit function and a curriculum-based training scheme, addressing exploration and credit assignment challenges.

Findings

CM3 learns faster than existing algorithms.

Effective in complex multi-goal multi-agent environments.

Demonstrates success in navigation, traffic, and game scenarios.

Abstract

A variety of cooperative multi-agent control problems require agents to achieve individual goals while contributing to collective success. This multi-goal multi-agent setting poses difficulties for recent algorithms, which primarily target settings with a single global reward, due to two new challenges: efficient exploration for learning both individual goal attainment and cooperation for others' success, and credit-assignment for interactions between actions and goals of different agents. To address both challenges, we restructure the problem into a novel two-stage curriculum, in which single-agent goal attainment is learned prior to learning multi-agent cooperation, and we derive a new multi-goal multi-agent policy gradient with a credit function for localized credit assignment. We use a function augmentation scheme to bridge value and policy functions across the curriculum. The complete architecture, called CM3, learns significantly faster than direct adaptations of existing algorithms on three challenging multi-goal multi-agent problems: cooperative navigation in difficult formations, negotiating multi-vehicle lane changes in the SUMO traffic simulator, and strategic cooperation in a Checkers environment.