電磁場逆問題分析計算的快速全局優(yōu)化算法研究.pdf

1、洳；j碩士學位論文⑧論文題目電磁場逆問題分析計算的快速全局優(yōu)化算法研究作者姓名——韭—藍—————一指導教師——籃韭直—藍控——一——擔出正—茲攮——一學科(專業(yè))—蟲工理監(jiān)皇疆整盔—一所在學院——蟲墨工程蘭盛———一浙江大學碩士學位論文AbstractThankstotheadvancesinboththecomputerscienceandelectromagnetics，theinverseelectromagneticprobl

2、emhasbecomeatopicalsubjectinCompumfionalElectromagneticsinordertomeettheeverinereasingrequirem鋤臨indesignautomationsofelectromagneticdevicesBasedonacomprehensiveanalysisendintegrationofavmlableglobaloptimalalgorithms，this

3、dissertationmainlyfocusesonthestudyoffastglobaloptimalmethodsappliedtoinverseelectromagneticproblemsFirstlytheGeneticAlgorithm(GA)andtheParadeSwarmOptimization(PSo)methodsarestudiedDifferentimprovementsageproposedtodevel

4、oprobustoptimizersinsensesofbothconvergencespeedsendglobalsearchabilitiesTypicalmathematicfunctionsaresolvedtodemonstratetheeffectivenessendemcieIlcyoftheimprovedGAandPSOSecondlytodeVelopcomputationaUyefficientglobalopti

5、maloptin[1izerswiththemaingoalofreleasingtherequirementfortremendouslyheavycomputationrcsoBle船especiallytoreducethenumberoffunctioncallsthatinvolvecomputationallyheavyproceduresinthestudyofelectromagneticinverseproblems，

6、theresponsesurfacemodels(RSM)OrmethodologiesareincorporatedintothestochasticglobaloptimalmethodsForthispurpose，thecompactlysupportedradialbasisfunctionandthemovingleastsquaresapproximationbasedonesareintegratedintotheafo

7、rementionedtwoglobaloptimizersAsaresulttwonovelhybridalgorithmsareintroducedFinallytheproposedhybridalgorithmsareappliedsuccessfullytostudybenchmarkinverseelectromagneticproblemsThenumericalresultsofTEAMWorkshopProblem22

8、endProblem25嬲reportedvalidatethefeasibilityandtherobustnessofthecorrespondingalgorithmsKeywords：FastglobaloptimalalgorithminverseelectromagneticproblemGeneticAlgorithmParticleSwarmOptimizationresponsesurfacemodeloptimald