隨機需求與多重商品供應鏈網路穩健最佳化之設計

A Robust Optimization Approach to Multiple-product Supply Chain Network Design Problem with Stochastic Customer Demands

供應鏈網路設計問題會受到許多不確定性因素影響，包括需求、價格、成本等。傳統供應鏈網路設計問題多假設市場需求為已知明確值，且多以營運成本最小化或營運利潤最大化作為供應鏈網路設計之目標式。忽略市場需求不確定性特質及其可能導致的營運利潤變動風險。故本研究應用穩健規劃理論，探討隨機性需求下之多個製造商、多個物流中心位址選擇暨多重商品配銷規劃問題，以決定最佳工廠、物流中心之位址及工廠各商品之生產量。本模式以總營運成本期望值、營運成本變異風險成本及供需失衡風險成本之加權值最小化為目標式，以確保所設計之供應鏈系統的穩健性。而供應系統總成本則包括工廠生產成本、運輸成本、物流中心營運成本及物流中心存貨成本等四項。本研究應用抽樣基礎式的樣本平均近似法求解此模式；最後，透過數值例分析與敏感度分析說明本模型之合理性與可應用性。

Many uncertain factors have an influence on a supply chain network design problem, including market demands, product prices, and costs. Many conventional approaches assume that market demands are known and deterministic. Furthermore, the objective functions of supply chain network design problems are mainly to minimize total costs or to maximize total profit. Uncertainty and induced risk of profit variations are often disregarded. In this study, we apply the robust optimization technique to explore a multiple-product supply chain network design problem with stochastic customer demands. The optimal locations of factories, locations of distribution centers, and production volume of each commodity of each factory are simultaneously determined in the proposed model. In order to assure the robustness of the supply chain network, the objective is to minimize a weighted sum of total expected costs, variability of total expected costs, and infeasibility penalties of system constraints. Herein, total expected cost is the sum of total production costs, total transportation costs, total operation costs of distribution centers, and total inventory costs of distribution centers. A sampling-based algorithm, sampling average approximation method, is adopted to find the heuristic solution of the proposed model. Finally, numerical examples are elaborated and sensitivity analyses are utilized for demonstrating the rationality and practicality of the proposed model.