多策略混合的花朵授粉算法  

Multi-strategies Flower Pollination Algorithm

作　　者：姚光磊 熊菊霞[1] 杨国武[2] 郑宏宇 YAO Guanglei;XIONG Juxia;YANG Guowu;ZHENG Hongyu(School of Mathermatics and Physics,Guangxi Minzu University,Nanning 530006,China;School of Computer Science and Engineering,University of Electronic Science and Technology of China,Chengdu 611731,China)

机构地区：[1]广西民族大学数理与物理学院,南宁530006 [2]电子科技大学计算机科学与工程学院,成都611731

出　　处：《小型微型计算机系统》2024年第3期613-620,共8页Journal of Chinese Computer Systems

基　　金：2022年广西科技基地和人才专项项目(桂科AD22080021)资助;2020年广西民族大学引进人才科研启动项目(2020KJQD05)资助;2019年广西自然科学基金面上项目(2019GXNSFAA18503)资助.

摘　　要：针对花朵授粉算法(FPA)在解决高维度问题时存在收敛速度慢和收敛精度低的问题,本文提出一种混合了多种策略的花朵授粉算法(MFPA).该算法通过使用自适应控制因子来动态地切换全局与局部搜索策略的使用;为了提高收敛速度和维持花粉种群多样性,提出一种基于多方信息的全局搜索策略;为了能探索到更充分的解空间,提出一种局部搜索策略;为提高算法的搜索解空间能力,引入特征选择策略降低问题复杂度.基于多种类型测试函数开展模拟实验,与多种优秀算法进行算法性能对比分析,实验结论:MFPA算法在收敛速度与精度方面有着更好的表现,适用于求解大规模复杂优化问题.Flower pollination algorithm(FPA)has the problem of slow convergence speed and low convergence accuracy when solving high-dimensional problems.In this paper,multi-strategies flower pollination algorithm(MFPA)is proposed.The algorithm controls the conversion between global exploration and local exploration dynamically by using adaptive control factors.A global exploration strategy for integrating multi-party information is proposed in order to increase the convergence speed and conserve the population polymorphic.A local exploration tactic is also proposed for searching more sufficient solution space.The feature selection tactic is introduced to decrease the difficulty of the issue solving and enhance the search ability of the algorithm effectively.A variety of test functions are used to carry out simulation experiments,and the algorithm performace is compared with that of the excellent algorithm.The experimental outcomes demonstrate that the MFPA has better performance in convergence speed and accuracy,and is suitable for solving large-scale difficult optimization problem.

关 键 词：花朵授粉算法 特征选择 多策略 最优解 自适应 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]