The emergence of on-demand passenger transportation services has disrupted the traditional transportation landscape, providing convenient and efficient mobility solutions. These services, facilitated by mobile applications, enable users to request rides at their convenience, connecting them with available drivers in real-time. This thesis studies on-demand mobility services in various capacities, from single-rider ride-sourcing service and ride-pooling service to on-demand transit, aiming to provide door-to-door convenience while exploring shareability. To facilitate the rapid growth and wide adoption of on-demand transportation, planning and operational strategies are essential.

These services, facilitated by mobile applications, have transformed urban mobility but are susceptible to v...[ Read more ]

The emergence of on-demand passenger transportation services has disrupted the traditional transportation landscape, providing convenient and efficient mobility solutions. These services, facilitated by mobile applications, enable users to request rides at their convenience, connecting them with available drivers in real-time. This thesis studies on-demand mobility services in various capacities, from single-rider ride-sourcing service and ride-pooling service to on-demand transit, aiming to provide door-to-door convenience while exploring shareability. To facilitate the rapid growth and wide adoption of on-demand transportation, planning and operational strategies are essential.

These services, facilitated by mobile applications, have transformed urban mobility but are susceptible to various uncertainties. This thesis addresses the challenges posed by uncertainties in these services. The study examines various variabilities from service time, pooling process, demand spatial distribution, and order generation and cancellation, examining their impact on service quality and performance metrics. The central question to be answered by this thesis is how to make robust decisions to improve the system’s performance in face of various uncertainties.

To answer the questions above, this thesis first discusses and compares two pricing strategies for a mixed ride-sourcing and ride-pooling market, investigating the effect of travel time variability. We develop a bi-level framework where the lower level investigates the modal split, zone choice and route choice equilibrium, and the upper level examines the optimal pricing strategy of the TNC company. We further apply this framework to a duopoly market to examine the influence of trip fare and cancellation penalty on waiting time, passenger demand and platform profits. Analytical results are obtained and pricing strategies are optimized for the platforms. For multi-capacity on-demand service, we design both the frequency-based flexible bus and dynamic routing systems by incorporating passenger arrival and cancellation uncertainties. Zonal based dial-a-ride problem is formulated to utilize the geographical adjacency of different orders. By proposing a reliability-based decomposition method, we improved the solution efficiency and implemented it for different real cases. This thesis provides insights into the current state of research and practical implications for on-demand passenger transportation under uncertainties. It serves as a foundation for future research in developing innovative strategies and technologies to enhance the resilience and efficiency of on-demand transportation systems in the face of uncertainties.