請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/6593
標題: | 軌道車輛運用路徑規劃尋優模式之建立 Development of Optimization Model for Rolling Stock Utilization Planning |
作者: | Kuo-Chu Liu 劉國著 |
指導教授: | 賴勇成 |
關鍵字: | 軌道運輸,運用路徑規劃,運用表,解構技術,變數產生法, Rail Transportation,Rolling Stock Utilization Planning,Utilization Schedule,Decomposition Skill,Column Generation Method, |
出版年 : | 2012 |
學位: | 碩士 |
摘要: | 軌道車輛為其營運所需中最昂貴的設備之一,如何有效的運用此項資產,以達到高效率的實務規劃,對整體系統的營運而言極為重要。本研究旨在探討車輛資源規劃中的「運用路徑規劃」問題,其係在考量路線、場站容量、車輛特性以及維修等諸多限制條件下,透過接續各車次及任務以生成「運用路徑」,再藉由多條運用路徑滿足欲規劃週期內所有車次的需求,而後彙整為「車輛運用表」於時刻表改點間重複使用。實務作業中,有經驗的作業員可透過人工排訂一套良好的車輛運用表,然其僅為一組近似解而無法保障其為全域最佳解。又車輛運用為實務中較細節的課題,過去的學術文獻涉獵較少,且多將問題大幅簡化,以致難以取得一套真正可輔助實務規劃使用的決策系統。
有鑑於此,本研究同時透過文獻回顧並對實務單位進行專家訪談,雙管齊下彙整影響運用路徑規劃決策的目標及因子。而後透過具有週期性的圓形二維時空網路圖作為問題架構,進而建立兩組不同的數學模式輔以求解。為了精進求解時間以為實務支援決策之用,本研究採用變數產生法中的Gilmore-Gomory演算流程,再將問題進行解構並發展一個「運用路徑規劃求解演算法」,透過有力的對偶理論為基礎,以建立一套軌道車輛運用路徑規劃的決策工具。 本研究透過求解臺鐵局實務案例,以驗證並測試其正確性與可行性。其規劃時間較人工而言改善許多,在未考量後續人員規劃的情況下,可維繫最低的迴送成本,並有效降低現有的檢修次數,使總成本可減少約20~25%。研究成果貼近實務並具有顯著績效,適合用以延伸推廣,以作為輔助實務作業員從事運用路徑規劃決策所用。 Rolling stock is one of the most expensive assets of a railway agency or company. Therefore, efficient utilization of rolling stock is a very important objective pursued in practice. This study proposed a Rolling Stock Utilization Planning Tool to improve the efficiency of rolling stock usage by creating a utilization schedule to cover the trips in the timetable with the consideration of practical requirements, such as inspection regulation, depot capacity, rolling stock characteristics, and most important of all, the idea of schedule circulation. Past studies usually simplified the problem by ignoring some of the important factors in practices; therefore, most of the railway agencies or companies still rely on experienced planners for this task. These practitioners can generally create an acceptable plan but there is no guarantee of the optimality. Consequently, there is a need for a planning tool which takes all important factors into account and provides optimal solutions to this problem. In order to develop such tool, this research first identified the appropriate factors by reviewing literatures and interviews with practitoners. Then the Rolling Stock Utilization Planning Model is developed to minimize the deadhead cost and inspection cost in rolling stock utilization planning cycle. This study also presents the column generation method with Gilmore-Gomory algorithm to improve the solution efficiency, especially for large-scale problem. The practical cases of Taiwan Railways Administration were used to demonstrate the validity and applicability of the proposed tool. The empirical results show that this optimization process can produce more efficient utilization plans with minimal deadhead distance and significant reduced number of inspections compared to the manual process. It’s equivalent to successfully reduce the total deadhead and inspection cost by 20%~25%. Using this decision support tool will help railways improve the efficiency of rolling stock utilization so as to provide reliable service to their customers. |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/6593 |
全文授權: | 同意授權(全球公開) |
顯示於系所單位: | 土木工程學系 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
ntu-101-1.pdf | 4.06 MB | Adobe PDF | 檢視/開啟 |
系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。