油罐車輛派遣決策支援系統建置之個案研究

Abstract

油罐車輛派遣之良莠影響油品配送的績效甚鉅，更是油品業者競爭致勝的關鍵。本研究以國內某家大型油品供應商爲個案對象，針對其油罐車輛作業特性，構建一套「油罐車輛派遣決策支援系統」，以提升油罐車輛派遣作業效率與公平性。本系統提供訂單輸入、油罐車輛派遣與車隊管理等功能。在核心的車輛派遣模組中，本研究以「委外車隊運送里程總和極小化」與「自有車隊運送里程差距極小化」爲目標，兼顧派遣之效率性與公平性，建構一個數學規劃模式並發展啟發式演算法來進行油罐車輛派遣。最後，本研究應用個案公司新竹運輸中心之實際營運資料（五種實例情境）進行系統測試與績效評估，結果顯示在委外車隊里程數方面，較原始派遣結果平均改善了37.8%；在自有車隊里程數差距方面，則較原始派遣結果平均改善了41.4%；操作時間方面約可節省57.1%的時間，效果相當顯著。

This paper aims to construct a decision support system (DSS), which is helpful in improving the performance of petroleum tank trucks dispatching. This DSS provides several functions, such as order entry, truck dispatching, fleet management, and friendly interface. We built a mathematical programming model and designed an effective heuristic method to dispatch trucks under two objectives: minimizing the overall mileages of contracted fleet, and minimizing the difference of mileages between private fleet. The computational results of five real-world scenarios implied that our heuristic performs significantly superior to the original dispatching results in both objectives.