Surrogate-assisted multi-objective covariance matrix adaptation evolution strategies

Open Webpage

Oladayo S. Ajani, Adeyinka Adedigba, Kalyana C. Veluvolu, Rammohan Mallipeddi

Published: 2025, Last Modified: 28 Sept 2025Appl. Soft Comput. 2025EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•An efficient offspring sampling scheme is proposed for multi-objective CMA-ES.•The sampling mechanism incorporates an ensemble of operators and a Gaussian Process (GP)-based surrogate model.•The performance of the proposed MO-CMA-EGO is evaluated on the WFG test suite and 18 Neural Architecture Search(NAS) problems.•Comparative analysis with other algorithms demonstrate the edge of MO-CMA-EGO.