离散二进制入侵杂草算法  

Invasive weed optimization algorithm of discrete binary version

作　　者：张帅[1] 王营冠[1] 夏凌楠[1] 

机构地区：[1]中国科学院上海微系统与信息技术研究所,上海200050

出　　处：《华中科技大学学报（自然科学版）》2011年第10期55-60,共6页Journal of Huazhong University of Science and Technology(Natural Science Edition)

基　　金：国家科技重大专项资助项目(2010ZX03006-002)

摘　　要：在继承入侵杂草优化(IWO)算法特性的基础上,提出了一种可用于解决离散空间组合优化问题的二进制入侵杂草优化(BIWO)算法.该算法保留了IWO算法的正态空间扩散特性,设计了一个扩散范围到扩散概率的映射函数,以概率的形式决定二进制杂草比特位的取值.选取连续空间的4个经典函数和离散空间的背包问题,对BIWO算法进行测试,结果表明:BIWO算法在寻优能力上性能优异,与二进制粒子群优化(BPSO)算法相比,BIWO算法更适合于离散空间的背包问题,能最终收敛到全局最优解或准最优解,且其寻到的最终解在均值和方差上均显著优于BPSO算法.On the basis of optimization invasive weed optimization（IWO） algorithm,IWO algorithm of binary version（BIWO） was proposed to solve the discrete combinatorial optimization problems.The method preserved the normal space diffusion feature of IWO algorithm,and designed a mapping function.This function could map the diffusion range into a probability,and the bits of binary weed changed its value according to the probability.Four classical functions in continuous space and the knapsack problems in discrete were used to test the BIWO algorithm.Experiment shows that BIWO algorithm works well in the ability of finding optimal value.Comparing with the binary particle swarm optimization（BPSO） algorithm,BIWO algorithm is more adaptable to the knapsack problems in discrete,which can finally get the global optimal solution or the second best solution,and the final solutions found by BIWO algorithm show better mean and variance than that of BPSO algorithm.

关 键 词：仿生算法 最优化 二进制序列 入侵杂草优化算法 背包问题 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]