Behavior-based neuroevolutionary training in reinforcement learning

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.

Combining Evolution and Deep Reinforcement Learning for Policy Search: a Survey

“…The bnet algorithm (Stork et al, 2021) is borderline in this survey as it does not truly uses an RL algorithm, but uses a Behavior-Based Neuroevolution (BBNE) mechanism which is only loosely inspired from RL algorithms, without relying on gradient descent. bnet combines a robust selection method based on standard fitness, a second mechanism based on the advantage of the behavior of an agent, and a third mechanism based on a surrogate estimate of the return of policies.…”