生產配銷系統在於決定最佳工廠和倉庫位置、規模、數量和配送的路線，因此在物流網路是非常重要的議題。設計此系統配置問題最常見為間斷型混整數規劃，但須花費許多時間收集詳細相關資料，及利用演算法求得最佳解，因此我們利用連續型近似模型求解，此方法只須藉由簡要的資訊和分析型模型。此篇論文主要探討，一間公司將在銷售市場設置新廠房以服務當地的顧客，然而每間工廠有各自的服務區域，產品在有限產能工廠生產後直接配送到當地顧客，且顧客的需求為不確定性。此篇論文目標為求得公司所需最小成本以符合顧客需求，且經由此模型可得知所需工廠數及主要運送的區域面積。最後，此篇提出數值範例和敏感度分析，以利經營者做決策。The production-distribution system design is one of the most important issues for the logistic network. It involves the determination of the location and size of the service facility and warehouse, transportation routes. Production-distribution system problems are solved usually by discrete mixed integer programming models which take much time to collect detailed data and obtain the optimal solution. But continuum approximation models need only concise summaries of data and analytic models. So we use continuous models to solve this problem. In this thesis, a firm wants to construct new facilities serve an entire market region where each facility serves a single sub-region. Items of customers’ demand are produced by the capacitated facility and are conveyed from the facility to customers by vehicles. Moreover, the customers’ demand is uncertain. Our goal is to find the minimum total cost for the production-distribution design problem. We can determine that how many facilities the firm needs and then get the size of the primary delivery region. Finally, we perform numerical studies and find how the parameters affect our models to facilitate decision makers.
