https://scholars.lib.ntu.edu.tw/handle/123456789/169142
Title: | 鐵路終端車站內列車調度及路徑指派問題之研究 | Authors: | 周義華 | Keywords: | 鐵路終端車站;列車運行計畫;列車調度;基因演算法;railway terminal station;train operation planning;train routing;genetic algorithms | Issue Date: | 2004 | Publisher: | 臺北市:國立臺灣大學土木工程學系暨研究所 | Abstract: | 要維持鐵路系統營運的水準須制 定鐵路列車運行計畫,包括:時刻運 行計畫、軌道使用計畫、車輛排程計 畫、乘務員排程計畫。雖然時刻運行 計畫可排除列車於車站間之運行衝突 ,但此階段並不考量車站的容量限制 ,為了確保列車能依據時刻表準時進 出車站,須進一步透過系統之調度原 則與方法,考慮車站內列車調度及路 徑指派的情形,安排各列車於車站內 之移動路徑與其停等佔用軌道之時間 ,其目的在於防止列車於車站內發生 運行衝突及確保列車能依時刻運行計 畫運行。 目前國外有關車站內列車調度問 題之研究多侷限於一般通過車站,且 不考慮列車在車站內的調度路徑指派 ,然而在終端車站中,列車之調度與 車場作業密不可分,車站與車場間之 列車移動頻繁,因此本研究擬同時考 慮終端車站內通過列車與進出車場列 車之調度與路徑指派問題,在避免運 行衝突,減少列車總延誤時間及增加 軌道使用效率的前提下,以數學規劃 中之節點包裹(Node Packing)模式 構建路徑指派問題,此模式經過適當 簡化後亦適用於一般通過車站之路徑 指派問題;在模式求解方面,由於使 用傳統正確解法之求解手續繁複,且 求解時間隨問題規模呈不確定之多項 式方式遞增,因此本研究擬透過啟發 式演算法中尋優解能力極佳之基因演 算法來求取最適解。希望透過模式求 解的程序,減少人工試誤所需時間, 提供一個明確且高效率的解決方案, 以利鐵路事業單位進行相關作業時參 考。 經過本研究利用台鐵松山車站與 南港客車場之實際資料進行實證後, 發現具多種啟發式運算子之基因演算 法求解品質極佳,而求解時間亦在合 理可接受之範圍內,且本模式可透過 不同目標層級之參數設定架構反映各 種調度人員規劃時考慮之因素,有效 滿足實際需求。 Train operation planning includes Time Tabling, Train Routing, Vehicle Scheduling, and Train Driver Scheduling. Time Tabling neglects the detailed layout of the railway network within the railway stations. Therefore, it may happen that a timetable is feasible with respect to the railway network between the railway stations but turn out to be infeasible if one also considers the detailed layout of the railway network within the railway station. The aim of Train Routing is to assist the planners in checking whether a timetable generated by Time Tabling is feasible with respect to the routing of the trains through the railway station. Most studies of Routing Trains through railway station considered only single station or yard. Since the routing of the trains within the railway yard influences the routing possibilities for the terminal station, a simultaneous determination for the routing of terminal station and yard may be desirable. At the prerequisite of minimum train delay and maximum track usage, we formulate the problem by “weighted node packing problem”. This formulation can also be used on intermediate station through modification. In order to solve the problem efficiently, we adopt the genetic algorithms rather than traditional optimization algorithm. By providing this problem-solving procedure, we seek to reduce the heavy burden of the planners. Through the empirical study, it was found that the solutions solved by genetic algorithms with heuristic operators had good qualities. In addition, to adjust the parameters of the objective function can also reflect the practical constraint and demand. |
URI: | http://ntur.lib.ntu.edu.tw//handle/246246/2859 | Other Identifiers: | 922211E002090 | Rights: | 國立臺灣大學土木工程學系暨研究所 |
Appears in Collections: | 土木工程學系 |
File | Description | Size | Format | |
---|---|---|---|---|
922211E002090.pdf | 96.18 kB | Adobe PDF | View/Open |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.