Landscape Analysis for Surrogate Models in the Evolutionary Black-Box Context

TL;DR

This paper explores how the features of a problem's landscape influence the accuracy of surrogate models in black-box optimization, using extensive analysis on benchmark functions with surrogate-assisted evolutionary strategies.

Contribution

It provides a detailed landscape analysis relating features to surrogate model accuracy, considering various transformations and data selection methods in black-box optimization.

Findings

Landscape features significantly impact surrogate model accuracy.

Transformations and data selection methods alter landscape feature effectiveness.

Analysis on benchmark functions demonstrates practical insights for surrogate modeling.

Abstract

Surrogate modeling has become a valuable technique for black-box optimization tasks with expensive evaluation of the objective function. In this paper, we investigate the relationship between the predictive accuracy of surrogate models and features of the black-box function landscape. We also study properties of features for landscape analysis in the context of different transformations and ways of selecting the input data. We perform the landscape analysis of a large set of data generated using runs of a surrogate-assisted version of the Covariance Matrix Adaptation Evolution Strategy on the noiseless part of the Comparing Continuous Optimisers benchmark function testbed.