混合和声搜索算法求解流水制造单元调度问题

Hybrid Harmony Search for Flowline Manufacturing Cell Scheduling

研究了单元制造系统(CMS)中,零件族准备时间与顺序相关(SDFSTs)的流水制造单元调度问题(FMCSP),求解目标为最大完工时间(makespan)最小化。根据问题特点提出了混合和声搜索算法(HHS),它通过迭代优化算子来提升解的质量,同时应用一个简单的发散策略来避免算法陷入早熟。该算法与当前最好的启发式算法和几个元启发式算法相比,在查找最优解方面优于这些算法,适合解决实际生产中的流水制造单元调度问题。

This paper considers the flowline manufacturing cell scheduling problem with sequence dependent family setup times （SDFSTs） in cellular manufacturing system（CMS） to minimize makespan. Based on the characteristics of this problem, a hybrid harmony search （HHS） algorithm is proposed. It uses an iterative optimization algorithm to enhance the quality of the solution and applies a simple diversification method to avoid the algorithm from conver- ging quickly. The proposed HHS algorithm is compared with the previous best heuristic algo- rithm and some state-of-the-art meta-heuristic algorithms on the same benchmark instances. HHS algorithm outperforms the existing heuristics in terms of finding a better makespan schedule. Thus, the proposed algorithm is quite useful for solving FMCSP with SDFSTs in practice.