Benchmarking CMA-ES with margin on the bbob-mixint testbed

Ryoki Hamano, Shota Saito, Masahiro Nomura, Shinichi Shirakawa

Abstract: The CMA-ES with Margin (CMA-ESwM) is a CMA-ES variant recently proposed for mixed-integer black-box optimization (MI-BBO), which introduces a lower bound on the marginal probability associated with integer variables. The CMA-ESwM shows promising performance compared to existing methods on simple benchmark functions. However, its performance has not been comprehensively investigated in other function classes, such as multimodal ones. In this work, we investigate the performance of the CMA-ESwM on the bbob-mixint testbed that includes problems of various properties for MI-BBO. The experimental results show that the CMA-ESwM outperforms the other MI-BBO methods at higher dimensions. The performance at low dimensions is competitive with the comparative methods.