Transfer Time Optimization in Transit Scheduling

This dissertation lays out new formulations of synchronized timetables for the bus timetabling problem to enhance the quality of transit service and transit ridership.First, a comparative analysis of five models is conducted in a deterministic setting by investigating model outputs at the network level (three nodes under different headway policies) and for each transfer node individually. A new model is introduced to relax the assumption of available vehicle capacity, an essential but neglected feature in most previous studies. The results indicate considerable implications for model outputs when demand is incorporated into the objective function and vehicle capacity is included in the formulation. Furthermore, it was found that agencies should take into account the location, demand distribution, and headway combination of transfer nodes while selecting the optimal transfer optimization model. Second, a timetable synchronization model is proposed incorporating bus dwell time determination which has been largely disregarded in the literature. A new concept of pre-planned holding time is also introduced to reduce the transfer waiting time for transfers to low-frequency routes while accounting for the penalty of extra in-vehicle time for onboard passengers and the possible consequences on headway regularity of a route. A Lagrangian relaxation-based heuristic is developed to obtain high-quality solutions efficiently. The experiments with up to 12 transfer nodes in the City of Toronto indicate that incorporating transfer holding time, dwell time determination, and vehicle capacity limit improves model outcomes considerably. Lastly, a stochastic optimization model is proposed considering variability in passenger walking times between bus stops at the transfer node, bus running times, dwell times, and demand uncertainty. The objective function includes transfer waiting times, delay times, and unnecessary in-vehicle times. The model determines dwell time by considering passenger arrival patterns at bus stops which have been neglected in previous transfer synchronization and timetabling models. A sample average approximation of the model is solved using a problem-based scenario reduction approach, and the Progressive Hedging algorithm. The experiments on two single transfer nodes in the City of Toronto demonstrate the potential advantages of incorporating stochasticity in transfer-based timetabling models and the high performance of the solution method.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.