Modelling For Park-and-Ride Network Design In A Bi-Modal Transportation Network

Park-and-Ride is an innovative travel mode which aims to mitigate this problem. Commuters use private car to finish their first part of the trip. When they are approaching the congestion area of the city, they park their car near a public transport station and transfer to the public travel mode to arrive their destination. This travel mode not only mitigate the traffic congestion in the CBD area of a city, but also alleviate the parking problem in metropolis.This paper adopt bi-level programming method to optimize the design of location and size of a Park-and-Ride scheme. The upper-level programming model reflect the traffic planner’s benefit which is mitigate the traffic congestion in CBD area; promoting the public transport usage. The lower-level programming model indicate the commuters’travel behavior which is the selection of travel mode and the selection of travel path. This paper first establish BNL-UE model to evaluate the mode split and route choice of commuters. This model can be solved with the excess-demand method after necessary network transformation. Due to this solution algorithm does not need to consume much storage in computers, it can be widely applied in practical engineering. The upper-level programming model is aiming to minimize the total travel cost in the network and promote the public transport usage.Based on the BNL-UE, this paper further developed a MNL-UE to better describe the mode split part of this combined model. The proof of equilibrium and uniqueness also be conducted in this paper. This model’s solution algorithm was developed based on the algorithm proposed by Ortuzar. This paper use goal programming approach to deal with the multi-objective phenomenon in upper-level programming model and developed a tabu search algorithm as the solution algorithm to solve this goal programming model.The chapter 4 of this paper studied the overlap of different transport mode and adopted CNL to mitigate the IID property of MNL model. The CNL-UE model was established. The equilibrium and uniqueness of the model was also explored. The DSD algorithm was used to solve it. For the upper level programming model, the chance constraint goal programming approach was adopt.