|Ph.D Student||Oliker Nurit|
|Subject||Transit Network Design that Considers Online Information|
|Department||Department of Civil and Environmental Engineering||Supervisor||Professor Shlomo Bekhor|
|Full Thesis text|
This research investigates the transit route network design problem. This problem has been researched in the literature. It consists of finding the best combination of transit routes and frequencies. The study deals with the different elements of the problem: route set generation, design optimization and the assignment model.
In practice, planning of transit networks is largely based on practical guidelines and designer intuition. As the mathematical problem is highly complex, most models developed in the literature are not applicable for large scale networks. Therefore, the main objective of the research is to develop an efficient solution model for the problem. Another main contribution of this research is the consideration of online information and strict capacity constraints in the assignment model, investigating passenger distribution in the network for different cases of available information.
The study develops a frequency based transit assignment model that consider online information of predicted arrival times is available to passengers, and accounts for strict capacity constraints. The study first includes both capacity constraints and online information in a frequency based transit assignment model. The model foresees a significantly different distribution of passenger in the transit network, compared to the case of no available information. The results demonstrate the potential impact of available information on travel behavior and the importance of its consideration in assignment models.
An infeasible start heuristic is developed for the transit network design. The method first assigns all candidate transit routes with the maximal frequency, and then iteratively eliminate routes and decrease frequencies of the less attractive ones. The running time of the method is very short compared to common meta-heuristic algorithms.
The transit network formed by the algorithm comprise fewer, faster and more frequent lines that serves high volume of passengers, compared to a given benchmark. The designed network serves the bulk demand well, reducing the average travel time in about 40% compared to the given transit network. The price of the significant reduction in the average travel time, is an increase in the walk only trip proportion. The potential benefit of not including a minimal demand satisfaction constraint in the design procedure should be noted. Possibly, the small share of demand can be served by a complimentary service.
The models developed in this thesis were applied using parameter values from the literature. Sensitivity analysis indicate that the model produces reasonable results. Further research will incorporate calibration of the model parameters in the assignment problem.