基于多策略改进的混合元模型全局优化算法及工程应用

打开文本图片集

中图分类号：TH123 文献标志码：A 文章编号：1671-5276（2025）06-0157-07

Abstract：Inordertoimprovehighlynonlinearandcomputation-intensiveengineeringoptimization，thispaperinitiatesanovel Multi-strategyenhanced hybridmetamodel-based globaloptimizationalgorithm（MS-HAM），which incorporatesahierarchical designspaceupdateapproachand variable sampling size strategies toenhance performance.Akeyregionupdate strategydivides thedesignspaceinto keyandcommonregions，andthekeyregionisadaptivelyupdatedbasedonsearchresultsto improve accuracyandconvergencespeed.Coarseanddensesamplingstrategiesareemployedatdiferentstagesofoptimizationtoenhance searcheficiencyandprevent prematureconvergencewithinalimiteddesignspace.Theefectivenessof theoptimizationalgorithm is validatedbycomparing itagainst existing methodsonthree benchmark optimization problemswithvarying numbersof variables.TheMS-HAMoptimizationmethod isaplied tooptimizethedesign parameters ofareal-worldengineeingcase study involvingaworkoverrigarm.TheresultsdemonstratethattheproposedMS-HAMalgorithm exhibitshighaccuracyandefficiency in optimizing computationally intensive problems，suggesting its broad applicability in engineering.

Keywords：MS-HAM algorithms；construction work vehicle；structural optimization

0 引言

随着科学技术的发展，在工程设计中，为了降低成本和时间，物理实验越来越广泛地被计算机模拟所取代[1]。（剩余8399字）