MMES: Mixture Model-Based Evolution Strategy for Large-Scale Optimization

This work provides an efficient sampling method for the covariance matrix adaptation evolution strategy (CMA-ES) in large-scale settings. In contract to Gaussian CMA-ES, proposed generates mutation vectors from a mixture model, which facilitates exploiting rich variable correlations of problem landscape within limited time budget. We analyze probability distribution this model and show that it approximates CMA-ES with controllable accuracy. use method, coupled novel strength adaptation, formulate based (MMES) -- variant optimization. The numerical simulations that, while significantly reducing complexity MMES preserves rotational invariance, is scalable high dimensional problems, competitive against state-of-the-arts performing global