学科分类
/ 1
1 个结果
  • 简介:Acceleratingtheconvergencespeedandavoidingthelocaloptimalsolutionaretwomaingoalsofparticleswarmoptimization(PSO).TheverybasicPSOmodelandsomevariantsofPSOdonotconsidertheenhancementoftheexplorativecapabilityofeachparticle.Thusthesemethodshaveaslowconvergencespeedandmaytrapintoalocaloptimalsolution.Toenhancetheexplorativecapabilityofparticles,aschemecalledexplorativecapabilityenhancementinPSO(ECE-PSO)isproposedbyintroducingsomevirtualparticlesinrandomdirectionswithrandomamplitude.Thelinearlydecreasingmethodrelatedtothemaximumiterationandthenonlinearlydecreasingmethodrelatedtothefitnessvalueofthegloballybestparticleareemployedtoproducevirtualparticles.TheabovetwomethodsarethoroughlycomparedwithfourrepresentativeadvancedPSOvariantsoneightunimodalandmultimodalbenchmarkproblems.ExperimentalresultsindicatethattheconvergencespeedandsolutionqualityofECE-PSOoutperformthestate-of-the-artPSOvariants.

  • 标签: 粒子群优化算法 探索能力 PSO算法 局部最优解 收敛速度 随机方向