Niche Radius Adaptation in the CMA-ES Niching Algorithm

摘要

Following the introduction of two niching methods within Evolution Strategies (ES), which have been presented recently and have been successfully applied to theoretical high-dimensional test functions, as well as to a real-life high-dimensional physics problem, the purpose of this study is to address the so-called niche radius problem. A new concept of adaptive individual niche radius, introduced here for the first time, is applied to the ES Niching with Covariance Matrix Adaptation (CMA) method. The proposed method is described in detail, and then tested on high-dimensional theoretical test functions. It is shown to be robust and to achieve satisfying results.