供应链模式预测控制林平和化学工程与材料工程系摘要一供应链系统.pdf

教專研 98P-008 供應鏈模式預測控制 供應鏈模式預測控制 林平和 化學工程與材料工程系 摘要 一供應鏈系統(supply chain system)可視為是一具有前置時間 (lead time)之線性離散系 統，當此系統在運作時，會發生所謂的長鞭效應(bullwhip effect) ，即在一供應鏈系統中， 即使顧客需求變異性不大，但當需求的資訊往上游傳遞時 ，便會一層一層被扭曲誇大， 造成越上游的廠商誤以為其下 游的客戶需求變異性非常大 。在供應鏈中長鞭效應雖然非 常特殊一種現象，但確是無法避免的。因而常使廠商做出 不正確的訂貨策略或錯誤的生 產規劃，使得廠商蒙受極大的生意損失。本文的目的就是提出一些控制方式來降低長鞭 效應的影響，因此應用數學之 z-轉換形成轉移函數以便於解決與分析供應鏈系統。在探 討供應鏈系統之動態行為時，所使用的控 制模式中，將發現結合模式預測控制(model predictive control)與前饋控制(feed-forward control)之模式，不僅在降低長鞭效應的效率最 好，而且在存貨的控管上也非常理想。 關鍵詞：供應鏈、長鞭效應、前饋控制、模式預測控制 Abstract A supply chain system can be viewed as a linear discrete system with lead times and operating constraints. The distortion of demand in upstream activities is known as “bullwhip effect＂, i.e. magnification of amplitudes of demand perturbations from the tail to upstream levels of the supply chain. The bullwhip effect is a special and inevitable phenomenon in a supply chain system. This paper will present several heuristic control laws for the model to solve effectively this phenomenon. Explicit transfer function model of the closed loop response is obtained by z-transform, so the entire chain can be modeled by connecting these transfer functions into a block diagram. Next we will use the system process control theory to handle the dynamic behaviors of a supply chain system. Among these control modes, we design a near-perfect control mode, i.e. feed-forward with model predictive control (MPC), to manage the supply chain system. The results reveal if using this control law we can find at least three cr