Optimize a Supply Chain Schedule under Variable Cost Functions

Abstract

在現今企業電子化的競爭環境中，利用供應鏈管理資訊系統處理客戶訂單和快速回應市場變動，已是必要且不可或缺的，然而，供應鏈管理資訊系統是否能有效地協助上、中、下游廠商安排最佳化的供應鏈排程，則完全取決於系統中供應鏈管理計量模式的有效性。尤其是當上、中、下游協同過程中，任何部份些微的儲運或生產活動的改變，均會牽動整個供應鏈的行程，因此，要能有效與迅速地安排最可行與最佳化的供應鏈排程，經理人除了溝通協調上的努力外，一個能切實反應真實儲運現況的供應鏈管理計量模式，更扮演著舉足輕重的角色。有鑑於目前市面上供應鏈管理資訊系統的計量模式，在處理成本函數時，不是將其簡化爲固定的數字或線性的函數、就是以啟發式或針對特定目地的演繹法來處理非線性成本函數，前者不符實際產業的運作，後者則不易被撰寫於一般供應鏈管理資訊系統中，且所計算出來的解答也只是局部最佳解。因此，本研究在於提出一些易於處理各色各樣非線性成本函數的方法，以便能讓供應鏈管理計量模式在目前的資訊系統中，更能發揮其強大與有效地供應鏈排程最佳化能力。一個小型的實例運算，則被用來展現所提出方法的實用性與廣泛性。

In current e-business competitive climate, employing supply chain management (SCM) information systems to fulfill customer orders and quickly response the changing market is obligatory and crucial. However, the effectiveness of computerized SCM system heavily depends on the availability of SCM quantitative formulation. Particularly, a simple change within the supply chain can lead to the complicated dynamics along the whole supply channels. Consequently, to effectively and efficiently arranging the most feasible and promising supply chain schedule requires not only arduous managerial efforts, but also a practical SCM quantitative model. Unfortunately, current mathematical models embedded in prevailing SCM systems either use fixed values or linear functions to replace real-world cost functions or employ heuristics or special purpose algorithms to treat nonlinear cost functions. The former approach is out of the reality, while the latter is hard to be coded in general commercial SCM systems and can not guarantee the computed result is global optimal solution. Accordingly, this work attempts to present some easy-to-handle formulations of treating various types of cost functions such as nonlinear step cost functions, concave cost functions, and S-curve cost functions, and to make current computer-based SCM quantitative models more practical to enhance the execution of more elaborate coordination activities. A small example is employed to demonstrate the practicality and applicability of the proposed methods.