一种改进的和声搜索算法求解FJSP

IMPROVED HARMONY SEARCH ALGORITHM FOR SOLVING FJSP

针对柔性作业车间调度问题(Flexible Job-Shop Scheduling Problem,FJSP),以最大完工时间为优化目标,提出一种新改进的和声搜索(Harmony Search,HS)算法。该算法采用两段组合的编码方式,利用全局随机混合初始化的方式来保证初始解的质量和多样性。算法的搜索过程中采用一次创作多个和声方式,充分利用和声记忆库积累的资源,来提高收敛速度。加入智能变异算子,达到均衡机器负载分配,提高搜索到最优解概率的目的。通过与文献算法对比基准测试结果和实际工程测试,证明该算法不但提高了求解FJSP的精确度,还具有较高的有效性和实用性。

For the flexible job shop scheduling problem(FJSP),we take the maximum completion time as the optimization objective and propose a new improved harmony search(HS)algorithm.The algorithm adopted the coding method with two-stage combination,and used the global random mixed initialization method to ensure the quality and diversity of initial solutions.In the search process of the algorithm,multiple harmony modes were created at one time,and the resources accumulated in the harmony memory database were made full use of to improve the convergence speed.An intelligent mutation operator was added to balance the load distribution of the machine,which improved the probability of finding the optimal solution.Compared with the literature algorithm,the benchmark test results and practical engineering tests show that the algorithm improves the accuracy of solving FJSP and has high effectiveness and practicability.