|Abstract: ||摘要:道路瓶頸之形成原因不外乎地形之限制(如隧道、橋樑等)及人為因素(如車道之被迫減少等)兩種。尤其在早上通勤時段中，因為大部份之通勤者希望到達工作場所的時間大致上相差不多，所以在通勤道路上之瓶頸路段入口前，排隊等候通過之車輛大排長龍之現象在每個工作天都會發生。由於通勤車輛排隊擁擠之情形最易發生於瓶頸路段，故擁擠收費政策在該背景下顯得格外重要。瓶頸路段擁擠收費之理論模式由Vickrey W. S. 在1969 年首先提出，但真正蓬勃發展乃是從1990 年開始，到目前為止這股研究熱潮都還在持續中。在這期間已有相當數量之論文陸續刊登於著名之國際學術期刊中，並且已成為探討道路定價之學術領域中的研究主流。在瓶頸路段擁擠收費之所有理論模式中，只有計畫主持人所發展之最佳階梯式擁擠收費模式推導出擁擠收費實施後可能帶來之排隊擁擠減少效果，以及探討擁擠收費實施後所有通勤者之出發行為。但該結果僅建立在線性模式之前提下，而不適用於非線性模式。這是理論模式中不足之部分，同時也造成決策者於事前評估階梯式擁擠收費實施後可能帶來之影響時產生不確定性，最後可能導致明顯的評估誤差，造成實施後與預期之結果有某種程度以上之落差，而形成整體社會成本之浪費。在所有可能之非線性函數中，僅有β分配是唯一符合階梯式擁擠收費模式之基本假設，故本研究以β分配作為建立非線性模式之主體架構，進而推導出擁擠收費實施後之排隊擁擠減少效果、通勤者出發分散之結果與出發時刻變動之過程。這些均衡結果將與線性模式下之已知均衡結果作一比較，以了解兩種不同模式下所有推導結果之差異何在。本研究之成果不僅在該學術領域的理論創新上有所貢獻，亦可提供決策者於擁擠收費實施前一些相當有參考價值之資訊，以利評估作業之順利進行。|
Abstract:Because of topographic restraints, such as tunnels and bridges, or artificial restraints, such as lane-decreasing regulation, commuting roads often have a narrow segment. A bottleneck is formed in these situations. Especially during the morning rush hours, because most commuters have almost the same preferred arrival time at work, there are often long and persistent queues in front of the entry to a bottleneck. Since queuing developed most frequently at a road bottleneck, the congestion pricing policy has become important under this circumstance. Vickrey W. S. (1969) first developed the model of pricing a queuing bottleneck, but the golden age of this field begins in 1990 and still popular in Transportation Economics nowadays. There are many research papers that belong to this field published in the famous academic journals. Also, this field has become the main research tendency towards the road pricing theory. Among the all models of pricing a queuing bottleneck, only the optimal step toll model, developed by Laih, provides analyses of the effects of the congestion toll on queuing reduction and on commuter departure behavior. But, unfortunately, these results are in effect only under the linear model, and the non-linear model is excluded. This not only makes the step toll model incomplete, but also creates uncertainties for policy makers in order to evaluate the step toll scheme before implementation. These uncertainties may result in evaluation errors, and finally lead to reverse the expected results after implementing the step toll scheme. Consequently, the social cost spent on the step toll scheme possibly become wasteful. Because only .beta.-distribution among the non-linear function satisfies with the basic assumptions for the step toll model, this research project adopts the .beta.-distribution as the main structure for the non-linear step toll model. Based on this model, we derive the effects of the optimal step toll on queuing reduction, departure decentralization and departure time switching decisions. Compare these equilibrium results with those in the linear model, the differences in all equilibrium results between the two models can be realized. These outcomes not only make the academic contribution in this research field, but also provide some valuable information for policy makers to evaluate the optimal step toll scheme.