Behavior-based Neuroevolutionary Training in Reinforcement Learning

TL;DR

This paper introduces a hybrid neuroevolutionary and value-based reinforcement learning algorithm that improves sample efficiency and learning speed by exploiting stored experiences and behavior modeling.

Contribution

It presents a novel hybrid approach combining neuroevolution with value-based RL, utilizing behavior-based loss functions and directed search in behavior space.

Findings

Enhanced sample efficiency over traditional evolutionary methods

Faster learning speed demonstrated on benchmarks and real-world problem

Effective behavior modeling improves policy optimization

Abstract

In addition to their undisputed success in solving classical optimization problems, neuroevolutionary and population-based algorithms have become an alternative to standard reinforcement learning methods. However, evolutionary methods often lack the sample efficiency of standard value-based methods that leverage gathered state and value experience. If reinforcement learning for real-world problems with significant resource cost is considered, sample efficiency is essential. The enhancement of evolutionary algorithms with experience exploiting methods is thus desired and promises valuable insights. This work presents a hybrid algorithm that combines topology-changing neuroevolutionary optimization with value-based reinforcement learning. We illustrate how the behavior of policies can be used to create distance and loss functions, which benefit from stored experiences and calculated state values. They allow us to model behavior and perform a directed search in the behavior space by gradient-free evolutionary algorithms and surrogate-based optimization. For this purpose, we consolidate different methods to generate and optimize agent policies, creating a diverse population. We exemplify the performance of our algorithm on standard benchmarks and a purpose-built real-world problem. Our results indicate that combining methods can enhance the sample efficiency and learning speed for evolutionary approaches.