高维0-1背包问题的双种群角度调制DE算法  

Angle modulation differential evolution with dual population for high dimension 0-1 knapsack problem

作　　者：邓长寿[1,2] 

机构地区：[1]九江学院信息科学与技术学院,江西九江332005 [2]合肥工业大学计算机网络系统研究所,合肥230009

出　　处：《计算机工程与应用》2010年第24期45-47,共3页Computer Engineering and Applications

基　　金：国家自然科学基金(No.50705039);江西省教育厅科技项目(No.GG10616)~~

摘　　要：针对高维0-1背包问题,提出一种双种群新型DE算法。该算法采用双种群编码机制,其中一个为低维的实数编码种群,另一个为高维的二进制编码种群。借鉴通信领域的角度调制原理,通过低维种群中的个体,生成高维种群个体,实现将高维优化问题转换到低维空间进行优化求解。此外,新定义丢弃算子对演化过程中的不可行解实时进行修正。仿真实验结果表明了该算法求解高维0-1背包问题的有效性。A novel differential evolution algorithm with dual population is proposed to solve the zero-one knapsack problems with high dimension.In the new algorithm,two populations are used during the evolution,with one float coding population and the other binary coding population.The angle modulation in the field of communication engineering is imported to generate high dimensional binary population with the low dimensional float coding population.In this way,the optimization problem with high dimension can be transformed into the low dimension space.Additionally,a new discarding operator is defined to fix up the infeasible solution.The results of two numerical experiments with different size show it is an effective way for the high dimension zero-one knapsack problems.

关 键 词：高维0-1背包问题 差异演化算法 双种群 角度调制 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]