一种改进的头脑风暴优化算法

针对头脑风暴优化（BSO）算法精度较差、后期收敛速度慢的问题，提出了一种改进的BSO（MBSO） 算法. MBSO算法通过对种群分组策略概率参数的调节，改变个体生成方式调节参与全局和局部搜索的个体比 例，算法前期加强全局搜索后期加强局部搜索，有效避免陷入局部最优.同时MBSO算法根据搜索所处不同阶 段采用可变最大步长的策略加速算法收敛并提高了优化精度.采用6个标准测试函数对MBSO算法搜索性能进 行了测试，与原始BSO算法、粒子群优化（PSO）算法和差分进化(DE）算法结果进行比较实验.仿真结果表 明，MBSO算法可以有效地避免陷入局部最优，稳定地找到更好的最优值，收敛速度显著加快. MBSO算法在 优化问题中表现出了优异的性能和巨大的潜力.
Aimingatsolvingtheproblemofpooraccuracyandslowconvergenceinlaterperiodofbrainstormingoptimi? zation(BSO)algorithm,amodifiedBSO(MBSO)algorithmisproposedbyadjustingtheprobabilityparameterofthepopu? lationgroupingstrategyandbyusingavariablemaximumstep.Thenewpopulationgroupstrategystrengthenedtheglob? alsearchintheearlystage,andenhancedthelocalsearchinthelaterstage,whicheffectivelyavoidedthesearchbeing trappedinthelocaloptimum.Atthesametime,theMBSOalgorithmusedvariablemaximumsteplengthaccordingtothe differentstagesofthealgorithm,whichacceleratedtheconvergenceandimprovedtheoptimizationaccuracy.Theexperi? mentalresultsof6typicalbenchmarkfunctionsshowedthattheproposedalgorithmissuperiortotheoriginalBSO,parti? cleswarmoptimization(PSO)algorithmanddifferentialevolution(DE)algorithminsolvingthelocaloptimumproblems, convergingtobetteroptimalvalues,andreducingthesearchtime.TheexcellentperformanceoftheMBSOalgorithm showsitsgreatpotentialin optimizationproblems.