摘要
分析了现有遗传算法在解决车间作业调度问题时局限产生的原因 ,提出了一种既能加快进化速度 ,又能提高抗早熟能力的嫁接遗传算法 ,并将其成功应用于车间作业调度问题。最后的实例测试表明了新算法的有效性和优越性以及它在调度领域的应用可行性。
The reason why genetic algorithm available exhibits limitations while applied to Job-Shop Scheduling Problem(JSSP) is analyzed, and an improved algorithm called grafted genetic algorithm featuring rapid speed of evolution and its strong ability to avoid premature convergence is presented. The convergence rates are speeded up and the premature convergences are decreased by the introduction of grafted population to direct the evolution-wanted population. Furthermore, the introduction of the crossover probability matrix can further raise the ability to avoid the premature convergence. Hence comes a breakthrough in soling two contradictive problems of GA. Finally, the validity of the proposed grafted genetic algorithm and the feasibility of its application in JSSP are illustrated with classic examples.
出处
《机械科学与技术》
CSCD
北大核心
2003年第6期873-875,878,共4页
Mechanical Science and Technology for Aerospace Engineering
关键词
嫁接遗传算法
车间作业调度
混合优化策略
Grafted genetic algorithm
Job-shop scheduling problem
Hybrid optimization strategy
