國道客運轉運區位、路線與排班之最佳化模式

Abstract

本研究目的主要整合國道客運軸輻式路網之模式，包含區位選擇、路網規劃及排班設計，以提升國道客運服務之效率，使乘客候車成本、乘客轉乘懲罰成本、業者營運成本，以及公共轉運站設置營運成本之總和最小化。 首先列舉出公共轉運站設置個數及區位之組合，其次，利用基因演算法將各組合產生之潛在營運路線進行路線分組，以決定路線間之轉運關係，及利用數學解析法進行路線之排班設計，求解各路線之最佳路網型式、最佳發車班距及設置公共轉運站之最佳區位。 本研究以政府之觀點來探討現行國道客運營運狀態，以行駛中山高速公路之國道客運國道營運路線為實例應用。列舉之設置公共轉運站個數及區位組合中，以設置公共轉運站於高速公路台北及台中交流道之組合總系統成本最小，因此，以該組合作為理想國道客運軸輻式路網，經轉運設計後，有154條營運路線，其中包含92條直達路線及62條轉運路線，總成本降低10%，顯示轉運之營運方式確能改善既有國道客運路網之經營。由敏感度分析知，乘客轉乘懲罰成本及乘客需求量係影響轉運系統之建立成功與否的關鍵，因此若降低乘客轉乘懲罰成本，並妥善設計轉運站之轉乘方式，以減少乘客轉乘之不便，更能發揮轉運系統建立之效果；乘客需求量與總成本呈正比，表示當乘客需求量增加時，總成本亦隨之增加，當乘客需求量減少時，總成本隨之減少，表示乘客需求量越低者，實施轉運之效果越大。

To further improve the efficiency of freeway bus service, this study proposes an integrated hub-and-spoke freeway bus networking model, including hub location, routing and scheduling, so as to minimize the total cost of government (hub cost), operators (operation cost), and passengers (waiting, traveling and transfer cost). Three methods are used to solve the proposed three-level problem: The total enumeration method is used to determine the optimal hub location under a given number of hubs (at most two hubs). Genetic algorithms (GAs) are then used to solve the routing problem under the pre-determined hub location. The analytical method is used to determine the optimal bus frequency for each of hub-and-spoke networking routes based on the pre-determined hub locations and routing. All three-level problems are simultaneously solved to ensure the optimality. In other words, to select the optimal hub location should be given that the optimal routing and scheduling for each of hub location perturbations. To demonstrate the applicability of the proposed model, a case study on freeway bus routes on Freeway No.1 is conducted. Based on the origin-destination passenger demand, the optimal result shows that a total of 154 bus routes, including 92 direct service routes and 62 transfer service routes, are desgined and the best location of hubs are at Taipei and Taichung interchanges. Comparing to the current direct service, the optimal network can curtail total cost by 10%. The sensitivity analysis indicates that passenger demand passenger transfer cost and waiting cost are the key factors in designing the hub-and-spoke