Evolutionary Reinforcement Learning via Cooperative Coevolution

TL;DR

This paper introduces CoERL, a cooperative coevolutionary approach to reinforcement learning that decomposes high-dimensional policy optimization into subproblems, improving efficiency and scalability over existing methods.

Contribution

The paper proposes a novel cooperative coevolutionary reinforcement learning algorithm that enhances scalability and efficiency by decomposing policies and directly searching for partial gradients.

Findings

CoERL outperforms seven state-of-the-art algorithms on six benchmark tasks.

Decomposition into subproblems improves scalability in high-dimensional spaces.

Ablation study confirms the effectiveness of CoERL's core components.

Abstract

Recently, evolutionary reinforcement learning has obtained much attention in various domains. Maintaining a population of actors, evolutionary reinforcement learning utilises the collected experiences to improve the behaviour policy through efficient exploration. However, the poor scalability of genetic operators limits the efficiency of optimising high-dimensional neural networks.To address this issue, this paper proposes a novel cooperative coevolutionary reinforcement learning (CoERL) algorithm. Inspired by cooperative coevolution, CoERL periodically and adaptively decomposes the policy optimisation problem into multiple subproblems and evolves a population of neural networks for each of the subproblems. Instead of using genetic operators, CoERL directly searches for partial gradients to update the policy. Updating policy with partial gradients maintains consistency between the behaviour spaces of parents and offspring across generations.The experiences collected by the population are then used to improve the entire policy, which enhances the sampling efficiency.Experiments on six benchmark locomotion tasks demonstrate that CoERL outperforms seven state-of-the-art algorithms and baselines.Ablation study verifies the unique contribution of CoERL's core ingredients.