多期運輸規劃問題

Abstract

本研究以典型運輸問題（Traditional Transportation Problem）為基，考量多個期間的供需變化情形並結合中期生產計畫（Aggregate Planning）的動態運輸問題，透過時空網路擴張方法（Space-Time Network Expansion Method）的概念發展多期運輸規劃問題（Multi-period Transportation Programming Model）並建立數學模式。 研究內容將提出多期運輸規劃問題的數學模式，並且以線性和簡單非線性規劃作為多期運輸規劃問題的求解法。此外將探討工廠產能預備水準、運輸途徑可供選擇的種類以及各期差異的運輸成本，如何影響多期運輸規劃的求解結果。最後就不同需求分布狀況為主題，探討決策者如何在生產成本、運輸成本、貨物滯留產生的資金成本之間取捨，以決定最適合的產能預備水準和運輸方案。

The research of transportation problems with time-varying demand patterns are less to present in the past. This thesis proposes a new problem, which called multi-period transportation planning model, basing on traditional transportation problems and the concepts of aggregate planning, and presents the use of space-time network expansion method, which could transform the dynamic networks to time-independent ones. According to the model, a linear and a simple non-linear programming methods are proposed to solve this problem with the support of LINGO tool. In addition, three computational studies on capacity reserving levels, amount of transportation varieties, and transportation costs varying between routes, are performed, evaluating how these factors changes overall costs in planning results. Due to the fact that changing demands may drive the performance of multi-period transportation planning model, 23 kinds of demand patterns are presented and solved, with analysis results demonstrating the effectiveness of the MPTP model at each time-varying demand modes, to offer information and suggestions for decision makers that devoted to down-costing through trading-off between production costs, transportation costs, and interests within inventories and quantities-on-routes.