臨時性公車路網調整之多目標最佳化模式

MULTI-OBJECTIVE OPTIMIZATION OF TEMPORARY BUS NETWORK ADJUSTMENT

臨時性公車路網調整問題係指公車營運在面臨特殊事件或活動時，如工程施工所造成的道路封閉、大型活動的交通管制措施等，導致部分路段受影響而無法使原有公車路線正常營運下，於事前所規劃的臨時性路網調整，同時將調整路線事先公告讓民眾知道，並於事件或活動結束後恢復原營運路線。然過去做法經常以人為方式憑藉主觀經驗處理，缺乏一套有系統的規劃標準，且無同時考量路網調整前後的關係與變動及對既有乘客的影響，也因此常造成資源的浪費與既有乘客的流失。
針對臨時性公車路網調整問題，本研究建構一多目標混合整數線性規劃模式，求解兩個相互衝突的目標式，分別為最大化路網相似度與最小化標準化後旅行時間，達到同時兼顧路網調整前後的變動與對既有乘客的影響，又能考量乘客服務水準的路網調整方案。模式中使用集合分割法改寫，以增加問題求解的效率，同時採用增強型ε-限制式法求解所建構模式的Pareto最佳解，用以改善過去研究於目標式單位不一致、權重訂定，以及不同目標間權衡取捨等問題。
為評估本模式的路網改善成效，本研究藉由測試案例驗證模式的合理性與有效性，以及透過情境案例分析，探討不同封閉路段位置與不同公車路線數目對於路網調整之影響。分析顯示本模式的調整結果合理且效果良好，並可依不同目標函數間的權衡取捨找出最滿意的調整方案，此研究成果有利於規劃與決策者於臨時性公車路網調整策略上提供最適參考與建議。

Temporary bus network adjustment problem is an important issue in planning and operation of bus transport systems. When bus operation faces special events or activities, such as road closures during construction and traffic control measures for special activities, some road sections are affected and some original bus routes may not be able to provide services again. Planners need to adjust the bus network in advance, and announce the adjusted bus routes to people. After the events or activities come to an end, the bus network will be restored to its original state. The bus network adjustment is to reduce the business impacts of operators, and maintain the original service levels of passengers. However, some past practices were often used based on their subjective experiences, which lacks a systematic planning standard and fails to consider the relationship of before-and-after adjustment, as well as the impacts on existing passengers, often resulting in the waste of resources and loss of existing passengers.
Hence, the study proposes a multi-objective mixed integer linear programming model with two conflicting objectives, which maximizes the bus network similarity and minimizes the normalized travel time, so as to consider both the change of bus network and the influence on existing passengers. In addition, the model is reformulated as a set-partitioning problem to increase the efficiency of problem solving. Also, augmentedε-constraint method is used to solve for Pareto optimal solutions, which can improve the problems of past studies like the inconsistency of different objectives, the presume of weights, and the evaluation of trade-offs between different objectives.
To evaluate the effectiveness of bus network adjustment, the study verifies the reasonability and effectiveness of the model by test examples. Also, through scenario analysis, the impacts of different closed road sections and different number of bus routes are discussed. The results show that the proposed model is reasonable and effective. Also, decision-makers can have the most satisfactory adjustment plan based on their trade-offs between different objectives. The results of study can provide a reference for planners in making temporary bus network adjustment.