考量原料分配下之混合流線型生產排程

Abstract

隨著企業間的競爭日益激烈，生產排程與原料分配的結合對於降低成本和提高獲利有著顯著的幫助。雖然在混合流線型生產排程與原料分配的研究領域已有許多成果，但多數研究僅針對單一議題進行探討，將兩者同時納入考量的研究在近幾年才開始逐漸發展。 在機械加工領域中，一些大型工件在加工後會產生體積較大的餘料，這些餘料可供製造較小型的工件，取代原先用於該工件而需採購的新料，從而有效降低原料成本。而工件該選擇新料或餘料來加工，及餘料的最佳化分配以降低原料成本，則是本研究中的研究主題。本研究探討原料分配與混合流線型生產排程，考慮餘料再利用與最小化排程總延遲時間，以找出在有效利用原料情況下的最佳生產排程。這個問題包括兩階段製程的非等效平行機台，在第一階段製程結束後會產生加工餘料，而不同工件的加工順序會影響餘料的產生時間點，從而改變可使用的原料選項。 本研究首先建立適用於此問題的混合整數規劃模型，並根據模型的一些特性提出具有兩階段的基因演算法，此演算法在計算目標式的部分結合整數模型進行求解，以提高面對大規模問題時的可行性。最後通過不同規模問題的測試，探討了模型與演算法各自的優劣。

With the increasing competition among enterprises, the integration of production scheduling and material allocation has a significant impact on reducing costs and improving profitability. Although there have been many achievements in the research field of hybrid flow shop scheduling and material allocation, most studies focus on a single issue. Research that considers both aspects simultaneously has begun to develop in recent years. In the field of machining, large workpieces often produce substantial amounts of residual material after processing. This residual material can be used to manufacture smaller workpieces, thereby reducing the need to purchase new material and effectively lowering material costs. The focus of this study is on determining whether to use new material or residual material for machining and optimizing the allocation of residual material to minimize material costs. This study explores material allocation and hybrid flow shop scheduling, considering leftover reuse and minimizing total tardiness to find the optimal production schedule under effective material utilization. This problem involves a two-stage process with unrelated parallel machines. Residual material are generated after the first stage of processing, and the processing order of different workpieces affects the timing of residual material generation, thereby altering the available material options. This study first establishes a mixed integer programming model applicable to this problem. Based on some characteristics of the model, a two-stage genetic algorithm is proposed. This algorithm combines the integer model in the objective function calculation to improve feasibility when dealing with large-scale problems. Finally, through testing on problems of different scales, the advantages and disadvantages of the model and algorithm are explored.