New methodologies for limited-stop bus service design problems

Abstract

To enhance the efficacy of bus systems for the benefit of both operators and patrons, limited-stop services have been offered. However, noticeable lacunae persist within existing research on their design. Firstly, the impact of in-vehicle congestion on passengers' choice of route has not been integrated into existing models. Secondly, the prevalent strategies for implementing limited-stop services are fraught with drawbacks: the tactical planning strategy is less effective, while the dynamic stop-skipping strategy demands high computational resources. Lastly, studies about the fleet size and timetable remain unexplored in the context of design problems. Their effect on limited-stop service designs is unclear.

To address these research gaps, this thesis undertakes a comprehensive study on the design problem of limited-stop services. A pioneering bi-level optimization model that includes in-vehicle congestion considerations is formulated, considering user equilibrium at the lower level. Furthermore, a novel two-stage dynamic strategy is articulated, designed to circumvent the limitations inherent to both tactical planning and dynamic stop-skipping strategies. Additionally, this work contributes an optimization model to determine the optimal fleet size and timetable for bus systems providing limited-stop services. Meanwhile, three efficient and effective solution methods are developed to solve these three above models, respectively.