| 摘要: | 以往在學術界研究之網路問題以空間維度的探討居多，且針對網路之節線與節點作為主要研究對象，模型分析層面之參數變化調整以敏感度分析為主。有鑑於此，本研究導入時間維度概念，拓展網路問題，將空間面加上時間面，利用時空網路流動技巧建立最佳化模型，並研發出脆弱度分析，考量現實世界之時間維度存在價值，以運輸領域為例，運輸網路擁有尖離峰特性，並將分析對象從原本空間面的節線與節點，延伸到時空面的整體運輸系統，包含路線(arc)、場站(node)與運具(component)等等，不只是參數權重上的放大縮小，還可將某特定時點下所設計之節點或節線關閉，以系統最佳化之觀點得到運輸系統於不同時空下所產生的變化，使分析上能更精準的符合實務現況且在未來提供更佳的決策。
本研究研發脆弱度分析，並引用Yan & Chen (2002)之模式，其從長途客運業者立場，設計一整合旅次需求和班次供給之車隊排程模式，運用多重時空網路流動技巧建立模式，以定式旅客與車輛於時空中流動之情況，透過系統最佳化求得最佳之車隊排程與班次表以服務旅客。在求解方法上，不同於前研究以拉氏演算法求解，本研究利用C++程式語言配合數學規劃軟體CPLEX進行模式求解。最後為評估模式的實用績效，以國內一長途客運公司之營運資料進行實例測試，並針對不同變數與參數進行脆弱度分析，探討不同時空下，使用者所在乎之節線與節點對系統最佳化結果所產生的影響。
;Back in the days, the network problems were mostly discussed in the spatial dimension academically, and the main research objects were arcs and nodes. The parameter adjustment of the model was mainly based on sensitivity analysis. In view of this, this study introduces the concept of time-space dimension, expands the network problem, uses the time-space network flow technique to establish an optimization model, and develops vulnerability analysis. Considering the value of the real-world, such as the peak moment characteristics of the transportation field, this study extends the research objects from arcs and nodes to the transportation system, including the routes (arcs) and stations (nodes) and cars (components). It is not only the change of the parameter weights, but also makes nodes or arcs can be closed which designed at a certain time point, and the changes of the transportation system in different time and space can be obtained from the viewpoint of system optimization, so that the analysis can be more accurate and provide better decisions for users in the future.
This study develops vulnerability analysis and refers to Yan & Chen (2002), a scheduling model and a solution algorithm for inter-city bus carriers. From the standpoint of inter-city bus company, it designs a fleet scheduling model that integrates passenger trip demands and bus trip supplies, and uses multiple time-space network flow techniques to establish a model to optimize in time-space dimension, and get the best fleet schedule to serve passengers. About solution method, different from the previous research, the Lagrangian algorithm is used to solve the problem, this study uses the C++ programming language and the mathematical programming software CPLEX to solve the model. Finally, for the practical performance to evaluate model, the case study is regarding the operation data of inter-city bus company in Taiwan, and the vulnerability analysis is carried out for different parameters and variables to explore the influence of the changes in different time and space on the system optimization results. |
