A Novel Alternate Point-taking Strategy for Surrogate-Assisted Evolutionary Algorithm

The essence of surrogate model is a low-cost alternative, which mainly replaces the computationally heavy simulation process to reduce time cost consumed. In past two decades, this approximation based optimization method has made remarkable progress, and models are widely used in expensive analysis optimization. addition, with development technology, no longer simple substitute, but can drive new sample points join training on historical data, so as gradually approach global optimal solution problem. For problems, there many surrogates-assisted algorithm methods. However, selection great influence accuracy model. order obtain more accurate model, newly added should meet diversity criterion specified distance, at same time, corresponding strategies be adopted fully explore sparse regions, avoid falling into local phenomenon process. Therefore, an ensemble surrogates alternate point-taking strategy (APTS) proposed, hierarchical search framework designed, using different algorithms each stage. effectiveness APTS verified three benchmark examples dimensions compared several advanced results show that better robustness than other methods most test problems.