應用於少量多樣化生產的多階段排程方法

Abstract

在少量多樣化的生產排程系統中，由於市場需求不穩定、產品多樣性高，導致備料困難和整備時間增加，使得機台無法有效利用產能。然而，一個好的排程應兼顧設備利用率、生產品質與準時交貨。為此，本研究出一種解決非相關平行機台排程問題的方法，利用多階段分解的概念，克服少量多樣化生產排程系統的限制，包含順序相依的整備時間、到達時間和機台的特性，最小化訂單延遲時間。 為了能提升排程求解效率，本研究提出的多階段排程方法將複雜的排程問題拆解為若干子問題，分別為工單指派、單機台工單排序和多機台排程問題：首先求解線性規劃模型，使指派的工單在各個機台達到工作量平衡，接著利用排序法則和交換方法改善工單在單一機台上的工作順序，最後再藉由調度工單至不同機台的方式，得到近似原問題的最佳解。 本研究亦將多階段排程方法應用於紡織業染色製程資料集並與文獻中結合變數鄰近下降法與反覆貪婪演算法、蜂群演算法、基因演算法所得出之優化演算法做比較，證實本研究之多階段排程方法在改善遲交訂單數與遲交天數方面可以顯著優於上述排程方法，找到排程最佳解。綜合以上，考慮生產製造環境的不確定性，本研究的方法不僅可以彈性的因應多變的生產環境、快速調配生產資源，更能提高產品達交率與顧客滿意度。

Due to unstable market demand and high product diversity, material preparation will be difficult and setup times will increase, preventing capacity from being effectively utilized in a high-mix low-volume production scheduling system. A good schedule, on the other hand, should consider equipment utilization, production quality, and on-time delivery all at once. Therefore, this research proposes the multi-stage scheduling method for the problem of tardiness minimization on unrelated parallel machine scheduling with sequence-dependent setup times, release times, and processing constraints. To improve scheduling efficiency, the multi-stage scheduling method divides the complex scheduling problem into several sub-problem, including job allocation, single-machine job sequencing, and multi-machine scheduling. First, the linear programming model is used to allocate jobs and balance the workload of each machine. Then, sorting rules and exchange methods are used to improve the order of jobs on a single machine. Finally, an optimal solution that approximates the original problem is obtained by moving jobs to different machines. Moreover, the results show that the multi-stage scheduling method outperforms methods combining variable neighborhood decent with iterated greedy search, artificial bee colony, and genetic algorithm on the textile dyeing process dataset in terms of reducing the number of tardy orders and tardy days. This research considers the uncertainty of the manufacturing environment, allowing it to not only respond to changing production environments and quickly allocate production resources but also improve the order fill rate and customer satisfaction.