特殊尖峰需求下鐵路列車排程規劃之最佳化模式

Abstract

台灣地區西部走廊城際運輸型態特性中，鐵路運輸在連續假期扮演著尖峰疏運之功能與角色，尤以民俗假期疏運返鄉人潮為最。因此，如何因應特殊長假期之高需求強度現象，建立一套符合旅客需求與改善原有排程績效之列車排程機制，一直是學術與實務探討之重要課題。 鐵路旅客列車排程規劃問題，涉及旅客需求、停站規劃、列車排班等三個不同層次階段之規劃作業。本研究嘗試構建一列車排程規劃模式，將營運者與使用者的權益同時兼顧於排程規劃機制之中，並以列車載客營收值與旅客旅行時間節省效益最大化作為規劃目標。同時為能在排程規劃之決策過程裡合理反映旅客需求，乃在調節列車運輸服務規劃作業時，依據市場區隔理念，區隔長、中、短程不同旅運長度的旅客市場服務設計。基於此，本研究首先參酌國內外文獻，經擷取鐵路旅客之不同旅次長度組合，以敘述性偏好法構建旅客偏好模式，並選取旅行時間差距、發車時間與有無座位服務為效用函數決策變數，作為旅客選擇行為之研究基礎。其次，本研究為有效運用現有的鐵路資源，引用優先權觀念，在特殊假期尖峰時段優先指派高等級列車與長運程旅客，並發展出能處理多車種列車最佳班次數之排程規劃問題的啟發式求解程序。 本啟發式演算法在求解所構建的旅客需求列車排程分析模式時，係透過假期尖峰旅客偏好模式作為列車營運計畫之重要媒介，而以規劃目標值為新列車排程之尋優指標，求解出最佳列車車種組合、發車時點排程、列車停站數及其停靠站址。最後，以民國86年清明節連續假期台鐵資料進行個案研究，經分析結果顯示確實能求得一成效不錯之可行排程方案，確認本尋優程序可因應旅客需求特性，決定最佳列車發車時點及其最佳停站組合方式的排程問題，並得以有效提昇鐵路運輸的整體效益。 本研究建議未來在特殊假期時段，台鐵應對旅客運輸市場作有效區隔和取捨，並落實「長途旅客搭乘鐵路列車，短途旅客搭乘公路客運」之構想，以達連續假期運輸資源有效運用及合理分配之目的。

In Taiwan, extraordinary high travel demand always occurs in intercity travel market during long-weekend holidays, which continuously bothers both transportation operators and their customers. Developing a better strategy for train dispatching to deal with such a special need has long been an important subject for both research study and real world operation. In real word operation, railroad train dispatching plan must consider the requirements in three different aspects including customer preference, train stopping scheduling and train dispatching. The study would like to define a strategy to better utilize the Taiwan Railway Administration (TRA) existing resources, which can provide long-distance customers a higher service priority in long-weekend holidays. It is trying to gain the best benefit for the society according to the concept of transportation demand management. It means if no matter how many efforts were devoted the service quality is hardly to improve simply because of excessive demand, then the service providers needs to define their main service subject from the railroad’s passenger market. Based on the concept of different traveling time, different departure time, and the availability of seat which are used to develop the decision function for travellers in each category, the during long-weekend holidays was divided into three different categories according to the distance of travel. According to this idea, the study has proposed a railroad train-dispatching model, which has considered railroad travellers’ preference in long-weekend holidays. Under this modelling approach, long distance customers and higher class of trains are given a higher priority in dispatching arrangement. This model can be used to develop a whole day schedule including train dispatching departure times and stopping pattern for each train. A heuristic algorithm has been applied to develop this analytical model. Maximum the combination of operation income and customers’ travelling time saving has been the objective of the model. Under this given objective, the model can obtain the best plan of train dispatching for a long-weekend holiday. A ticket selling data provided by the TRA in long-weekend holidays during 1997 has been used to verify the model. Results indicate that the model does provide a better service than traditional ways. Customer time saving is the most obvious benefit provided by the model. It is recommended that under a special condition, such as long-weekend holidays, to better utilize the operation resources, each mode of transportation needs to identify its service market according to their special capability in transportation. Generally most transportation professionals believe that railroad must input more resources in long distance customer market, and bus companies should pay more attention on the short distance customer market. Ideally, this kind of arrangement is better off for the society in many aspects.