- 學位論文
考量人員排程之路邊租還電動共享汽車充電與調度最佳化
Optimization of Vehicle Charging and Relocation in Free-floating Electric Carsharing Systems Considering Crew Scheduling
摘要
由於私有車輛對於都市土地造成極大負擔並造成空氣汙染及交通壅塞，加上購置汽車需負擔高額的費用，因此共享汽車系統近年來在全球蓬勃發展。其中，路邊租還電動共享汽車系統以兼具環保及彈性的租賃方式，成為一項受歡迎的新興運輸服務。然而，旅次在空間及時間上分佈不平均的特性造成車輛分佈不均是營運者必會面臨的問題，因此營運者往往需要透過適當的調度策略才能維持系統的服務水準。本研究目標以基於營運者的調度策略最大化營運者利潤，針對路邊租還電動共享汽車系統建構一個可求解車輛調度問題並考慮充電站及人員排程的最佳化模式。本研究首先建構一個雙層路網，分別記錄車輛及調度人員的流動，且充電站在路網設計中以一系列的複製節點設置；接著提出一個混合整數規劃模型，然而，由於數學規劃模型需要大量的求解時間，因此本研究另外套用了一個以搬運組為求解核心的啟發式演算法，並將有關充電站的調度行為納入考量。本研究將充電站節點加入搬運組的設計、進一步將搬運組區分為兩種類型，並探討不同類型搬運組被納入局部搜索或是相互替換的機制。藉由不同的案例測試及敏感度分析，結果顯示當車隊數量及調度人員的數量改變時，對於旅次方向較單一的情境影響較大；而充電站數量的改變則對於旅次方向較平均的情境影響較大。在求解時間上，旅次方向較單一的情境則普遍需要更多的時間。本研究可提供路邊租還電動共享汽車系統營運者關於車輛充電、調度與營運之決策參考，進而增加車輛使用效率並降低人員調度成本。
並列摘要
Since private cars have imposed great pressures on the urban land use, inducing the problems of air pollution and traffic congestion, and a car ownership is also accompanied with high expenses, carsharing system has been developed rapidly around the world. Free-floating carsharing (FFCS) system with electric vehicles (EVs) is an emerging transportation service with popularity due to the eco-friendly characteristic and the flexible rental style. However, imbalanced vehicle distribution resulted from temporally and spatially uneven demands of users is the problem that must be faced by the FFCS operators. To deal with the problem, the operator needs to conduct appropriate relocation strategies to maintain the level of service in the system. The aim of this research was to maximize the profit of the operator based on an operator-based relocation strategy. An optimization framework for solving the vehicle relocation problems in FFCS systems with considering charging stations and crew scheduling was proposed. A double-layered network was first built to record the flows of EVs and the crew respectively. In the network, a series of copied nodes for charging stations were set. Then, a Mixed-integer Programming (MIP) model was constructed. However, due to the considerable solving times of the MIP model, a heuristic algorithm basically developed on the concept of relocation pair was additionally applied, and the relocations regarding the charging stations were considered. Charging-station nodes were joined to the design of relocation pairs, and the relocation pairs were further classified into two types. The mechanisms for different types of relocation pairs being added into the searching process of the local search and the interchange were investigated. By testing different cases and conducting sensitive analysis, the results showed that the scenario of single-direction trips was more sensitive to the number of EVs and the number of crew members; while the scenario of multiple-direction trips was more sensitive to the number of charging stations. For the running time, the scenario of single-direction trips was generally longer. In conclude, this research provides the operators of FFCS systems a reference to establish operational strategies for vehicle charging, relocation and management which improve the efficiency of the vehicle operation and reduce the personnel relocation cost.