Using Automated Algorithm Configuration for Parameter Control

TL;DR

This paper introduces a new benchmark for dynamic algorithm configuration by controlling a key parameter in a genetic algorithm for OneMax problems, demonstrating improved performance through automated configuration techniques.

Contribution

It proposes a novel DAC benchmark based on controlling a parameter in a genetic algorithm and shows how static automated configuration can outperform existing policies.

Findings

Automated configuration significantly improves performance on the benchmark.

The landscape of parameter-control policies is characterized.

Stronger baseline policies are computed via numerical approximation.

Abstract

Dynamic Algorithm Configuration (DAC) tackles the question of how to automatically learn policies to control parameters of algorithms in a data-driven fashion. This question has received considerable attention from the evolutionary community in recent years. Having a good benchmark collection to gain structural understanding on the effectiveness and limitations of different solution methods for DAC is therefore strongly desirable. Following recent work on proposing DAC benchmarks with well-understood theoretical properties and ground truth information, in this work, we suggest as a new DAC benchmark the controlling of the key parameter $\lambda$ in the $(1+(\lambda,\lambda))$~Genetic Algorithm for solving OneMax problems. We conduct a study on how to solve the DAC problem via the use of (static) automated algorithm configuration on the benchmark, and propose techniques to significantly improve the performance of the approach. Our approach is able to consistently outperform the default parameter control policy of the benchmark derived from previous theoretical work on sufficiently large problem sizes. We also present new findings on the landscape of the parameter-control search policies and propose methods to compute stronger baselines for the benchmark via numerical approximations of the true optimal policies.