基于改进遗传算法的复杂产品DSM最优开发模块划分研究  

DSM Optimal Development Module Division of Complex Product Based on Improved Genetic Algorithm

作　　者：裴小兵[1] 赵璇子 王秀丽[1] PEI Xiaobing;ZHAO Xuanzi;WANG Xiuli(School of Management,Tianjin University of Technology,Tianjin 300384,China)

机构地区：[1]天津理工大学管理学院,天津300384

出　　处：《工业工程与管理》2021年第6期84-94,共11页Industrial Engineering and Management

基　　金：国家创新方法工作专项资助项目(2017IM060200)。

摘　　要：针对传统遗传算法求解DSM模块划分问题时,由于染色体结构固定的影响,最大模块划分数量受限,导致无法得到最优解,本文提出一种基于改进遗传算法的产品模块划分方法。在种群初始化过程中生成异构染色体种群;在传统遗传算法中插入变邻域搜索算法,设计多邻域结构,查找突破最大模块划分数量限制的最优模块划分结果;提出自适应的交叉和变异算子,可对异构染色体进行交叉和变异操作,保证了交叉变异的有效性和对优秀结构的留存。最后通过案例计算分析,验证该算法的可行性和有效性。When traditional genetic algorithm was used to solve DSM module division problem,the number of maximum module division was limited due to the fixed chromosome structure,which might lead to the defect that the optimal solution could not be obtained.Therefore,a product module partitioning method based on improved genetic algorithm was proposed.In the process of population initialization,heterogeneous chromosome population was generated.The variable neighborhood search algorithm was inserted into the traditional genetic algorithm.The multi-neighborhood structure could help to find the optimal module partition result without being restricted by the maximum number of modules.An adaptive crossover and mutation operator was proposed to perform crossover and mutation operations on heterogeneous chromosomes,ensuring the effectiveness of crossover and mutation and the retention of excellent structures.Finally,the feasibility and effectiveness of the algorithm are verified by case analysis.

关 键 词：设计结构矩阵 模块划分 遗传算法 变邻域搜索 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]