A review of surrogate-assisted evolutionary algorithms for expensive optimization problems

• Published in: Expert systems with applications Vol. 217; p. 119495
• Main Authors: He, Chunlin, Zhang, Yong, Gong, Dunwei, Ji, Xinfang
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.05.2023
• Subjects: Evolutionary algorithms; Expensive optimization; Surrogate model; Swarm intelligence; Expensive optimization; Swarm intelligence; Evolutionary algorithms; Surrogate model
• ISSN: 0957-4174; 1873-6793
• DOI: 10.1016/j.eswa.2022.119495

Many problems in real life can be seen as Expensive Optimization Problems (EOPs). Compared with traditional optimization problems, the evaluation cost of candidate solutions for EOPs is expensive and even unaffordable. Surrogate-assisted evolutionary algorithms (SAEAs) has become a hot technology to solve EOPs in recent year, because they can effectively reduce computational cost and improve solving efficiency. However, few literatures provide a systematic overview for SAEAs. This paper systematically summarizes the existing research results of SAEAs from the aspects of algorithms and applications. Firstly, the necessity of studying SAEAs and several commonly used surrogate models are introduced. Subsequently, according to the type of objective functions and constraints, the existing SAEAs are classified and discussed. Then, the application of SAEAs in many fields are reviewed. Finally, we indicate several promising lines of research that are worthy of devotion in future. •Existing surrogate-assisted evolutionary algorithms are classified.•Different surrogate-assisted evolutionary algorithms are analyzed.•Applications of surrogate-assisted evolutionary algorithms are reviewed.•Several promising research directions in the future are indicated.