面向装配序列规划的模因算法研究  

Research on Memetic Algorithm for Assembly Sequence Planning

作　　者：张金敏[1] 胡晓宇 Zhang Jinmin;Hu Xiaoyu(School of Mechanical Engineering,Lanzhou Jiaotong University,Lanzhou 730070,China)

机构地区：[1]兰州交通大学机电工程学院

出　　处：《机械科学与技术》2019年第11期1771-1777,共7页Mechanical Science and Technology for Aerospace Engineering

基　　金：国家自然科学基金项目(51375049)资助

摘　　要：针对遗传算法在求解装配序列规划问题中收敛速度慢、产生重复解等问题,提出一种基于模因算法的装配序列规划方法。将模因算法中全局搜索与局部搜索相结合动态更新种群的策略引入,采用装配优先约束矩阵和干涉矩阵建立装配规划模型,并以装配单元之间的相异性之和构建适应度函数。在非干涉解空间中进行全局搜索,获得装配规划方案,再通过二叉树中序遍历调序算法将较优方案转化为可行解。通过交叉操作和变异操作后,在可行解空间内进行局部搜索,获得较优解。通过典型柱塞油泵装配规划验证了该算法的可行性和可靠性;并将其与遗传算法进行比较,证明其更有效。To improve the general genetic algorithm for solving the assembly sequence planning(ASP) problem that it has slow search speed and massive repeated solutions, an ASP method based on the memetic algorithm(MA) was proposed. The strategy of dynamically updating the number of population by combining global search with local search is introduced. The assembly sequence planning model that adopts the constraint matrix and the interference matrix was constructed, and the fitness function for calculating the sum of similarity between assembly units was established. The assembly sequence planning is globally searched in the non-interference solution space. In order to traverse binary trees, the sorting algorithm is adopted to transform an optimal assembly sequence planning solution into a feasible solution. The optimal solution is locally searched in the feasible solution space through crossover and mutation operations. The assembly sequence planning process of a typical plunger pump is used as an example to prove the feasibility of the proposed ASP method. Compared with the genetic algorithm, the algorithm proposed in the paper is more effective.

关 键 词：装配序列规划 模因算法 遗传算法 适应度函数 

分 类 号：TH162[机械工程—机械制造及自动化] TP391[自动化与计算机技术—计算机应用技术]