基于独立种群遗传算法求解JSP问题  

Independent Population-Based Genetic Algorithm for JSP

作　　者：徐家旺[1] 

机构地区：[1]沈阳航空航天大学,沈阳322000

出　　处：《科技通报》2013年第4期136-138,共3页Bulletin of Science and Technology

基　　金：教育部人文社科规划基金(11YJA630165)

摘　　要：为解决传统单种群遗传算法在求解JSP问题上存在的搜索能力不强,计算结果质量不高,优化时间较长等问题,提出了一种基于独立种群的遗传算法。本文在分析了多种群遗传算法优势的基础上,根据车间调度问题的特点,设计了多种基于调度规则的初始化算法,并为保证种群进化过程的独立,提出了多种交叉、变异算子。最后给出了仿真算例,说明了多种群遗传算法的有效性。To solve the search capabilities of traditional single-population genetic algorithm for TSP,the calculation is not high quality,optimizing a long time and other issues,an new genetic algorithm based on independent populations was proposed.According to the characteristics of job shop scheduling problem,this paper designed a variety of initialization algorithm based on the scheduling rules,based on analysis of a variety of population genetic algorithm advantage,and then a variety of crossover,mutation operator was proposed to ensure the independence of the process of population evolution.Finally,a simulation example illustrated the effectiveness of a variety of population genetic algorithm.

关 键 词：多种群 遗传算法 车间调度 

分 类 号：TP163[自动化与计算机技术—控制理论与控制工程]