自适应多目标遗传算法在柔性工作车间调度中的应用  

Application of Self-adaptive Multi-objective Genetic Algorithm in Flexible Job Shop Scheduling

作　　者：陈华平[1] 谷峰[1] 卢冰原[1] 古春生[1] 

机构地区：[1]中国科学技术大学信息管理与决策科学系,合肥230026

出　　处：《系统仿真学报》2006年第8期2271-2274,2288,共5页Journal of System Simulation

基　　金：安徽省自然科学基金项目(050460404)

摘　　要：针对柔性工作车间调度问题的特点,提出了一种新的自适应多目标遗传算法,其特点包括:同时运用了基于工序顺序和基于机器分配两种交叉变异方法以弥补经典工作车间调度问题中交叉变异操作的局限性;根据遗传算法搜索的历史自适应的调整两种交叉变异方法的概率以提高算法的搜索效率和稳定性;引入多目标遗传算法中的小生境技术以保持种群的多样性;采用精英保留策略保护进化过程中的优秀个体。实验结果证明该算法在多目标柔性工作车间调度问题的应用上,可以产生分布性较好的高质量的解。According to the characteristic of flexible job shop scheduling, a new self-adaptive multi-objective genetic algorithm was proposed. It features the usages of crossover and mutations based on job order and machine assignment simultaneous in order to offset the crossover and mutations＇ localizations of traditional job shop scheduling, a self-adaptive adjustability of two crossover and mutations＇ probability based on the history of genetic search in order to improve search efficiency and stability, and a niching technique of multi-objective genetic algorithm in order to maintain the diversity of group. And the strategy of preserving excellent individual was also proposed. The experiments show the self-adaptive multi-objective genetic algorithm could produce well-distributed and high-quality solutions in multi-objective flexible job shop scheduling.

关 键 词：柔性工作车间调度 自适应 多目标 遗传算法 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]