Automated Configuration of Genetic Algorithms by Tuning for Anytime Performance

TL;DR

This paper compares hyperparameter tuning methods for genetic algorithms across diverse problems, highlighting the effectiveness of using anytime performance measures for configuration and analyzing sensitivity to budget allocation.

Contribution

It introduces a comparative analysis of automated configuration techniques for genetic algorithms and emphasizes the benefits of using anytime performance metrics.

Findings

Using the area under the empirical CDF curve often yields better configurations.

Tuning for expected running time is highly sensitive to the allocated budget.

Anytime performance measures can be more effective for configuration tasks.

Abstract

Finding the best configuration of algorithms' hyperparameters for a given optimization problem is an important task in evolutionary computation. We compare in this work the results of four different hyperparameter tuning approaches for a family of genetic algorithms on 25 diverse pseudo-Boolean optimization problems. More precisely, we compare previously obtained results from a grid search with those obtained from three automated configuration techniques: iterated racing, mixed-integer parallel efficient global optimization, and mixed-integer evolutionary strategies. Using two different cost metrics, expected running time and the area under the empirical cumulative distribution function curve, we find that in several cases the best configurations with respect to expected running time are obtained when using the area under the empirical cumulative distribution function curve as the cost metric during the configuration process. Our results suggest that even when interested in expected running time performance, it might be preferable to use anytime performance measures for the configuration task. We also observe that tuning for expected running time is much more sensitive with respect to the budget that is allocated to the target algorithms.