求解全局優(yōu)化問(wèn)題的正交協(xié)方差矩陣自適應(yīng)進(jìn)化策略算法

《求解全局優(yōu)化問(wèn)題的正交協(xié)方差矩陣自適應(yīng)進(jìn)化策略算法》由會(huì)員上傳分享，免費(fèi)在線閱讀，更多相關(guān)內(nèi)容在行業(yè)資料-天天文庫(kù)。

1、求解全局優(yōu)化問(wèn)題的正交協(xié)方差矩陣自適應(yīng)進(jìn)化策略算法摘要:針對(duì)協(xié)方差矩陣自適應(yīng)進(jìn)化策略(cmaes)求解高維多模態(tài)函數(shù)時(shí)存在早熟收斂及求解精度不高的缺陷,提出一種融合量化正交設(shè)計(jì)(od/q)思想的正交cmaes算法。首先利用小種群的cmaes進(jìn)行快速搜索,當(dāng)算法陷入局部極值時(shí),依據(jù)當(dāng)前最好解的位置動(dòng)態(tài)選取基向量,接著利用od/q構(gòu)造的試驗(yàn)向量探測(cè)包括極值附近區(qū)域在內(nèi)的整個(gè)搜索空間,從而引導(dǎo)算法跳出局部最優(yōu)。通過(guò)對(duì)6個(gè)高維多模態(tài)標(biāo)準(zhǔn)函數(shù)進(jìn)行測(cè)試并與其他算法相比較,其結(jié)果表明,正交cmaes算法具有更好的搜索精度、收斂速度和全局尋優(yōu)性能。關(guān)鍵詞:協(xié)方差矩陣自適應(yīng)

2、進(jìn)化策略;正交設(shè)計(jì);高維多模態(tài);進(jìn)化策略;函數(shù)優(yōu)化hybridorthogonalcmaesforsolvingglobaloptimizationproblemshuangya.fei1,2*,liangxi.ming1,chenyi.xiong11.schoolofinformationscienceandengineering,centralsouthuniversity,changshahunan410083,china;2.schoolofelectricandinformationengineering,changshau

3、niversityofscienceandtechnology,changshahunan410114,chinaabstract:inordertoovercometheshortcomingsofcovariancematrixadaptationevolutionstrategy(cmaes),suchasprematureconvergenceandlowprecision,whenitisusedinhigh-dimensionalmultimodaloptimization,anhybridalgorithmcombinedcmaeswitho

4、rthogonaldesignwithquantization(od/q)wasproposedinthisstudy.firstly,thesmallpopulationcmaeswasusedtorealizeafastsearching.whenorthogonalcmaesalgorithmtrappedinlocalextremum,basevectorsforod/qwereselecteddynamicallybasedonthepositionofcurrentbestsolution.thentheentiresolutionspace,

5、includingthefieldaroundextremevalue,wasexploredbytrialvectorsgeneratedbyod/q.theproposedalgorithmwasguidedbythisprocessjumpingoutofthelocaloptimum.thenewapproachistestedonsixhigh-dimensionalmultimodalbenchmarkfunctions.comparedwithotheralgorithms,thenewalgorithmhasbettersearchprec

6、ision,convergentspeedandcapacityofglobalsearch.inordertoovercometheshortcomingsofcovariancematrixadaptationevolutionstrategy(cmaes),suchasprematureconvergenceandlowprecision,whenitisusedinhigh.dimensionalmultimodaloptimization,ahybridalgorithmcombinedcmaeswithorthogonaldesignwithq

7、uantization(od/q)wasproposed.firstly,thesmallpopulationcmaeswasusedtorealizeafastsearching.whenorthogonalcmaesalgorithmtrappedinlocalextremum,basevectorsforod/qwereselecteddynamicallybasedonthepositionofcurrentbestsolution.thentheentiresolutionspace,includingthefieldaroundextremev

8、alue,wasexploredbytrialvectorsgen