基于自适应领导者选择与完全反向学习机制的混合花粉樽海鞘群算法

打开文本图片集

关键词：混合算法；樽海鞘群算法；花粉算法；领导者选择；完全反向学习中图分类号：TP301.6 文献标志码：A 文章编号：1001-3695（2026）04-015-1089-09doi：10.19734/j.issn.1001-3695.2025.08.0302

Hybrid salp swarm algorithm and flower pollination algorithm with adaptive leader selection and full opposition-based learning mechanism

Li Dahai，Wu Sikai†，Wang Zhendong （Schoolof InformationEngineering，Jiangxi UniversityofScienceand Technology，Ganzhou Jiangxi 341Ooo，China）

Abstract：To tackle the deficiencies such as insuficientconvergenceaccuracyandfrequent premature behaviorof a single heuristicalgorithm，this paperproposedtheHSF-ALSFOalgorithm.Atfirst，HSF-ALSFO hybridized thesalpswarmalgorithm （SSA）with thflowerpollnationalgorithm（FPA）.Itriedtotakeadvantageofboththehighlyeficient globalsearchcapabilityof SSAandtheexcelentlocalsearchcapabilityofFPA.Secondly，itadoptedanadaptiveleaderselectionmechanism， whichselectedeliteindividualsbyconsideringboththeirfitness valuesandrelativedistancesastheleaderstoguidethepositionupdatingof therestedidividuals.Atlast，itappiedthefullopposition-basedlearningwithrandom Givensrotationto generateoppositesofall individualsinthepopulationtoraiseupthediversityof thewhole population.ItusedCEC202 test suite asthe testbed to evaluate the performanceof HSF-ALSFO with the vanill SSA，suchas four refined SSA：AGSSA， EKSSA，TLSSA，DMSSA，and threeimproved FPA：SCFPA，EFPA，ABFPA.The resultsof Friedman test based on the experimentdatademonstrates that HSF-ALSFOcanexhibitasignificantdiferencecomparedwithco-evaluatedalgorithms.Ablation experimentsalsoillstratethat HSF-ALSFOcanattin theoptimal performanceunderthecombinationofallimprovement strategies.

Keywords：hybridalgorithm；salpswarmalgorithm；flowerpollnationalgorithm；leaderselection；fullopposition-based learning

0 引言

元启发式算法的概念及其发展源于对复杂优化问题的求解需求，尤其是当传统的精确算法（比如分支定界法、整数规划法等）无法适用时。（剩余20613字）