- 學位論文
運用基因演算法求解多目標低溫物流車輛途程問題
Multi-Objective Vehicle Routing Problem for Cold Chain Industry Using Genetic Algorithm
摘要
面對全球暖化日益嚴重，企業紛紛提出方案以減少碳排放量，就低溫物流業者而言，運輸工具往往是為耗費資源及破壞環境的主要來源之一。若能發展一套有效機制降低碳排放量，不僅減少對環境的傷害，也可減低企業營運成本，甚而提升企業的形象、建立競爭優勢。有鑑於此，本研究建構最少碳排放量及最短距離的多目標數學模式求解車輛途程問題。 本研究分為兩階段進行求解，第一階段運用k-means演算法進行分群；第二階段再運用基因演算法求解多目標車輛路徑規劃問題。為驗證模式之有效性，以台灣某知名低溫物流公司的情境及資料為基礎，運用基因演算法進行試驗，結果共分為三個部分。第一部分，僅考慮配送，探討距離與重量對車輛碳排放量的影響，先將較重的商品配送至物流據點，可降低路徑的總碳排放量；第二部分，車輛除了配送外，同時行駛至相鄰的集貨據點服務，能有效降低總碳排放量與總運輸距離；第三部分，車輛除了不超載外，也考慮車輛總運行時間上限，會導致車輛數量增加，致使總距離與總碳排放量增加的結果。針對以上不同測試結果，可作為決策者執行模式的參考依據。
並列摘要
In order to reduce the threatening effects of global warming, all enterprises must face and deal with the problem of carbon emission carefully. This study proposes a mechanism to solve the multi-objective vehicle routing problem for the cold chain industry. Both the total carbon emission and the total travel distance are minimized. A genetic algorithm (GA) is employed to find the solutions. The proposed mechanism is divided into two-stages. At the first stage, a k-means clustering algorithm is used to group customer locations. At the second stage, GA is employed to find out the minimized CO2 emissions and the minimized delivery routes with shortest travel distances. The proposed approach is demonstrated experimentally by a variety of scenarios of a famous cold chain company in Taiwan. The experimental results show that GA can find solutions effectively. In addition, the CO2 emissions according to a minimum emission objective in some cases may be less than those according to a shortest travel distance.