構建決策支援系統模擬物流派車策略之研究：以王屋科技為例

Abstract

近年來由於全球化的競爭日趨明顯，許多企業皆把全球化佈局列為企業發展重點，並積極改善整個企業流程相關的成本。企業的供應鏈包括產品設計、採購、倉儲、生產、運輸配送、行銷及售後服務等，如何在滿足顧客需求的前提下，以最低的成本，將產品有效率地配送至顧客的運輸問題是供應鏈中一個重要的課題。有鑑於運輸配送對於供應鏈的重要性，本研究以一個案公司為實例對象，針對物流系統中貨物配送的時機與配送路線的最佳化為目標來設計一套決策支援系統（DSS），並利用現有的資料來模擬訂單的發生，用以測試本研究所設計的配送策略的績效，包括配送的運輸成本與服務水準。 本研究的個案對象為王屋科技公司，主要的產品為燈俱用的變壓器，其位於廣東省珠江三角洲內，因河川眾多，收費站林立，致使運輸配送成本高居不下。本研究針對物流配送的議題，設計一套決策支援系統，包括四項模組：人機界面模組、派車排程決策模組、車輛路線決策模組與資料庫模組。其中，派車排程決策模組是用來決定各訂單的派車交貨日期；車輛路線決策模組是用來計算派車的最短路徑。 本研究以蒙地卡羅演算法來模擬訂單的發生，並以延遲(Postponement)的概念設計三項派車策略：併裝(Consolidation)、延遲(Postponement, 90%)與快速回應(Quick Responsive)，之後再以之前所設計的DSS模組來模擬求解這三項策略的績效，並分析其在成本與服務水準的得失(trade off)變化。 模擬分析後發現：Œ快速回應的策略有最好的回應能力，服務水準最高，但是因為配送次數頻繁，所以造成其運輸成本最高； 併裝的策略也就是完全延遲的策略有最低的運輸成本，因為此策略的重點為降低配送次數、增加承載率，但是其服務水準為最低：Ž延遲加上一經濟配送容量限制的策略的績效介於上述兩者中間。在需求小時，延遲策略會降低配送次數，使得運輸成本下降；在需求大時，延遲策略會適當的提前配送，使得服務水準增加。 與現況作比較時，採用延遲(90%)的策略可降低運輸成本61%，將使運輸成本由佔總成本的8%降至3.2%。

This research attempts to design the logistic dispatching strategies of supply chain, and discusses the trade-off between the transportation cost and the level of service. The analytical framework of this research is based on the decision support system and Monte Carlo simulation. In this thesis, we take Wang-House as the case for our study. This company’s main products are the transformer of lamps and lanterns. The half of it’s customers locate in China, and the others locate in Europe and America. It faces a problem of high transportation cost in China. First, we construct a decision support system (DSS). The DSS includes four main models: dialog interface/system control model, dispatching scheduling decision model, vehicle routing decision model and database management model. The dispatching scheduling decision model is to determine the dispatching date of each order. And the vehicle routing decision model is to determine the shortest path of dispatching. Second, we use the Monte Carlo simulation to simulate the real orders, and take advantage of these orders to test the performance of our dispatching strategies. Third, we use the concept of postponement to design three dispatching strategies: consolidation, postponement (90%) and quick response. After we have the simulate orders and making strategies, we can use the DSS models to test the results of different strategies. Finally, through the simulate tests, we have conclusions: 1.For the quick response strategy, it has the highest level of service. But it also has the highest transportation cost because the high frequency dispatching. 2.For the consolidation strategy, it has the lowest transportation cost, but it also has the lowest level of service. 3.For the postponement (90%) strategy, the results of this strategy are between quick response and consolidation. When the demands are low, this strategy can lower the transportation cost by reducing the dispatching frequency. When the demands are high, this strategy will dispatch in advance to increase the level of service. And the postponement (90%) strategy can reduce the transportation cost from 8% to 3.2%.