改进初始种群的遗传算法解决柔性车间调度  

Genetic Algorithm of Improved Initial Population for the Flexible Job Shop Scheduling Problem

作　　者：余智勇[1] 袁逸萍[1] 李晓娟[1] 

机构地区：[1]新疆大学机械工程学院,新疆乌鲁木齐830047

出　　处：《机械设计与制造》2016年第11期258-260,264,共4页Machinery Design & Manufacture

基　　金：新疆维吾尔自治区自然科学基金(2014211A008)

摘　　要：初始种群对遗传算法求解的质量和速度有决定性影响,传统遗传算法求解FJSP问题时,一般是随机生成初始种群,在迭代初期会形成许多无效方案,只有经过复杂的运算才会形成较优的方案,这样就会降低算法的收敛速度,对柔性作业车间调度的特点进行研究之后,提出了对初始种群给予基于全程检索规则编码生成初始种群的策略,提高初始种群质量的同时,也不会失去其多样性,而且还能提高全局收敛性。实例用改进的遗传算法,将结果与用传统遗传算法得到的结果比较,证明了改进算法的优势。The quality of initial population of genetic algorithm hove a decisive influence on the quality and the speed. When the traditional genetic algorithm is applied in solving flexible job shop scheduling problems, the initial population is randomly generated, which may result informing many infeasible solutions at the beginning of the iteration. Only through a complex operation will form optimum solutions, it may greatly reduces convergence speed of the algorithm.After study the characteristics of the flexible job shop scheduling,lnitial population give rules of base on the entire Search to code and generate initial population＇s strategy, has been put forward. When the quality of initial population be improved,its diversity also won＇t lose. At the same time,its global convergence can be improved.The instance in this article using the improved genetic algorithm,the results are compared with the traditional genetic algorithm＇s results.lt proved that the advantages of improved algorithm.

关 键 词：遗传算法 柔性作业车间调度 初始种群 

分 类 号：TH16[机械工程—机械制造及自动化] TH186