偏遠地區大眾運輸補貼分配與服務型態規劃之整合模式

Abstract

近年來台灣經濟成長造成道路交通日漸惡化，為因應此變化，交通部分別在民國八十五年及民國九十一年提出「促進大眾運輸發展方案」與「發展大眾運輸條例」以期改善道路交通環境。雖然都市因發展大眾運輸系統，交通獲得改善，但對於人口稀少以及運輸需求低落的偏遠地區，發展大眾運輸服務易\產生虧損及效率低落等問題。 本研究目的為提出整合最佳補貼分配與服務型態規劃之雙層數學規劃模式。上層模式為中央政府如何在總補貼預算限制下，將補貼預算分配給不同的地方政府；下層模式為地方政府根據所分配到的補貼預算，如何規劃其運輸服務達到大眾運輸使用率最大化。在上層模式中考慮到兩目標式:公平性(最小化地區間大眾運輸使用率差距)與效率(最大化大眾運輸使用率)。模式中考慮到不同的運輸服務型態:固定路線直達公車服務、固定路線幅軸路網公車服務與彈性路線公車服務。偏遠地區居民則利用羅吉特模式做為選擇大眾運輸服務與私人運具(汽車)的依據。為了簡化演算法來解決提出的問題，上層決策者被替換為幾個普遍被採用的補貼分配政策，而基因演算法則被用來求解下層問題(決定運輸服務的最佳組合方式)。為了確認模式適用性，本研究利用簡例進行分析。結果表明，根據不同的人口密度，不同偏遠地區擁有不同的最佳大眾運輸服務。除此之外，另外比較不同補貼分配策略下的公平性與績效。最後，進一步利用敏感度分析調查其重要參數對模式的影響。

The rapid economic growth in Taiwan leads to the severe deterioration of the level of service of road, and excess delays, energy consumption and emissions. In order to improve such a condition, the Ministry of Transportation and Communications (MOTC) implemented several public transportation improvement strategies, including The Alternatives for Promoting the Development of Public Transportation and The Statutes for Public Transportation Development during 1996-2002. Although the public transportation in some urban areas has been significantly improved, the public transportation in rural areas with sparsely population and low transport, demand till suffering from seriously operation deficit and low patronage problems. Based on this, this study proposes an integrated bi-level mathematic programming model for the optimal subsidy allocation and service planning for rural area transit service. The upper level assumes the subsidy is allocated by the central government to different local governments under a subsidy budget constraint; while the lower level is for local government to plan the optimal transit service for the rural areas in its jurisdiction based on the allocated subsidy to maximize its transit usage rate. Two objective functions in the upper level are considered: fairness (to minimize the differences in transit usage rate among areas) and efficiency (to maximize the aggregate transit usage rate). Different transit service types are considered: fixed-route direct bus service, fixed-route hub-and-spoke bus service and flexible bus service. Residents in rural area choose among these transit modes and private vehicles (cars) based on a mode choice Logit model. To simplify the solution algorithms to the proposed model, the decisions of the upper level are replaced by several commonly adopted subsidy allocation strategies while a genetic algorithm is used to solve the lower level problem (to determine the optimal combination of transit services). A case study on an exemplified example is conducted to examine the applicability of the proposed model. Results show that different optimal transit services are solved in different rural areas depending upon their population densities and the fairness and efficiency indices of different subsidy allocation strategies are also compared. At last, the effects of key parameters are further investigated by sensitive analyses.