摘要
传统遗传算法求解装配序列规划问题时会在初始化过程中产生大量非可行序列,影响求解速度并且导致最终得到的规划解质量不高。针对该问题,通过在初始化过程中加入启发式算子和基于无向图的广度优先搜索策略,保证了初始化个体的可行度和高适应度;在此基础上,对装配规划问题的特征重新对选择、交叉和变异算子进行了设计。实验表明,改进后的算法具有更好的稳定性和高效性,并且算法执行过程中不需要注入人工信息,使得装配过程更加自动化。
During the process of solving the assembly sequence planning problem,traditional genetic algorithm will generate a large number of useless initial assembly sequences which affect the solution rate and lead to low quality of planning solutions.Both a heuristic operator and a breadth-first search strategy,which is based on undirected graph,are added into to the initial process of genetic algorithm,so that the feasible degree and the fitness of individuals are ensured.Furthermore,according to the features of assembly sequence planning problem,operators including selection,crossover and mutation are re-designed.Experimental results show that the improved algorithm have better stability and efficiency compared with traditional genetic algorithm,and the algorithm implementation process need not any manual information,therefore the whole assembly process is more automated.
出处
《桂林电子科技大学学报》
2012年第2期129-133,共5页
Journal of Guilin University of Electronic Technology
基金
广西研究生教育创新计划(2011105950812M23)
关键词
遗传算法
装配序列规划
广度优先搜索策略
无向图
加权启发因子
genetic algorithm
assembly sequence planning
breadth-first search strategy
undirected graph
weighted heuristic factor
