Individual specialization in multi-task environments with multiagent reinforcement learners

TL;DR

This paper investigates how multi-agent reinforcement learning agents can develop specialization in multi-task environments, highlighting the limitations of epsilon-greedy exploration and proposing entropy-regularized policy methods for better convergence.

Contribution

It introduces a novel approach using entropy-regularized policy updates for independent learners, improving specialization and convergence in multi-task multi-agent settings.

Findings

Epsilon-greedy exploration synchronizes agents, hindering specialization.

Entropy-regularized policies lead to smoother convergence.

Increased number of agents correlates with higher likelihood of specialization.

Abstract

There is a growing interest in Multi-Agent Reinforcement Learning (MARL) as the first steps towards building general intelligent agents that learn to make low and high-level decisions in non-stationary complex environments in the presence of other agents. Previous results point us towards increased conditions for coordination, efficiency/fairness, and common-pool resource sharing. We further study coordination in multi-task environments where several rewarding tasks can be performed and thus agents don't necessarily need to perform well in all tasks, but under certain conditions may specialize. An observation derived from the study is that epsilon greedy exploration of value-based reinforcement learning methods is not adequate for multi-agent independent learners because the epsilon parameter that controls the probability of selecting a random action synchronizes the agents artificially and forces them to have deterministic policies at the same time. By using policy-based methods with independent entropy regularised exploration updates, we achieved a better and smoother convergence. Another result that needs to be further investigated is that with an increased number of agents specialization tends to be more probable.