面向昂贵优化问题的代理辅助进化算法综述OA

Many practical engineering problems can be categorized as expensive optimization problems(EOPs),charac-terized by high and sometimes prohibitive costs associated with evaluating candidate solutions.Surrogate-assisted evolu-tionary algorithms(SAEAs)have gained attention in recent years as a solution to EOPs due to their ability to reduce com-putational costs and increase solving efficiency.This paper provides a systematic overview of the research achievements in SAEAs from both algorithmic and application perspectives.It starts by explaining the necessity of studying SAEAs,then introduces several commonly used surrogate models.Next,it classifies and discusses existing SAEAs according to the problem types.Additionally,it summarizes current applications of SAEAs across various fields.Finally,it highlights the current challenges in SAEAs and offers insights into future trends and research directions in this field.