Research On The Bi-Level Model Of The Optimal Fare Structures Of The Urban Transit System

Under the national strategies of the " Bus Priority" and the " Giving Priority to Public Transport Development", it has become a hot topic to study on the optimal public transit fare policies to improve the effectiveness of prsent urban transit system. The urban transit fare structure and level are essential components of urban transit system, which affects the passengers’travel behavior and route choices. Unfair transit fare policeis would lead to the unbanlance of transit assignment and the improper distribution of transit resources. This thesis would take the optimal transit fare structure as its heart, and take the fare structure as its objective. A reasonable and practical transit fare pricing and subsidying model will be constructured, which will provide a new idea and method to transit fare decision process.This paper begins with the definition and the composition of urban transit system. The industrial properties of public transit are analyzed as follow, based on which the principles of transit pricing and subsidying are determined. After the discussion of the three main stakeholders, which are government, transit enterprises and passengers, in the process of transit pricing and subsidying, the concept of Game Theory is introduced to describe the relationship among them. Thirdly, the pros and cons of the most common transit fare structures are introduced and analyzed. Then, a new O-D based fare structure is introduced. Fourthly, a bi-level programming model is built to decribe the fare decision process. The upper-level is a priciple-agent game between government and transit enterprises. A path-based stochastic transit assignment (STA) model with elastic demand is first built to evaluate any given fare function.This thesis has the following research findings:the bi-level programming model and priciple-agnet game theory are utilized in describing the game between government, transit enterprises and pasengers; a path-based stochastic transit assignemnt is introduced, which takes the full consideration of the heterogeneousness of different transit lines under various fare structures; the common transit fare structures are analyzed in simple and real transit networks.