Efficiency analysis and optimal control of e-hailing service under demand surge after mass gathering events

Abstract

When mass gathering events (such as concerts or sports events) end, the demand for passenger flow in limited areas often increases sharply, which not only puts enormous pressure on the transportation system, but also brings significant safety risks. As an emerging mode of transportation, the introduction of e-hailing services not only enhances transportation capacity, but also provides a new approach for evacuating large passenger flows. However, despite a large amount of research in recent years focusing on the impact of e-hailing services on urban transportation, there is still little in-depth exploration of their application potential and control strategies in responding to large passenger flow events. In response to this research gap, innovative and systematic analysis has been conducted to improve the service efficiency of e-hailing services when large passenger flows happen. This thesis focuses on how to make e-hailing services efficiently and reasonably serve the evacuation of passengers in urban transportation systems. The following three problems are addressed, to improve e-hailing service's management under demand surge conditions. We first address the control problem of e-hailing vehicles in an evacuation system under demand surge case. A mathematical model with Trip-based Macroscopic Fundamental Diagram (MFD) representation is proposed to capture the dynamic passenger evacuation rates and traffic conditions within a bi-modal system. We adopt Logit model to determine passengers' mode choice between bus and e-hailing vehicles during evacuation. To accelerate passenger evacuation in such scenarios, we introduce two perimeter control strategies based on Proportional-Integral (PI) controller. These strategies regulate the total inflow rate of e-hailing vehicles and background traffic based on vehicle accumulation and the number of parked vehicles within the area, respectively. The second part investigates the pricing problem of e-hailing services in a two-region evacuation system. We improve the above Trip-based MFD model to illustrate the time-varying evolution of e-hailing vehicles within a two-region transportation system. The model takes into account the bilateral dynamics of drivers' repositioning decisions, passengers' demand dynamics, and dispatching between the two parties. To assist the e-hailing platform in ensuring a balanced distribution of e-hailing vehicle supply within a two-region transportation system, we propose a region-dependent pricing strategy and compare it with the undifferentiated static pricing strategy. This strategy adjusts e-hailing drivers’ per unit time wages in different regions, guiding passengers’ and drivers’ behaviors in the market effectively. Furthermore, we explore the design of boarding space in a bi-modal evacuation system, considering both efficiency and safety under demand surge case. The evacuation dynamics are captured utilizing three-dimensional macroscopic fundamental diagram (3D-MFD). We model the total generalized cost including users’ cost, operating cost and safety cost. Passengers’ mode choice between bus and e-hailing vehicles is also determined by Logit model. Two types of safety cost are mainly considered, including both pedestrian-vehicle conflicts and pedestrian-pedestrian conflicts. Four transport scenarios with different boarding space design are modeled and compared. The effectiveness of the models and control methods proposed for the three problems are all validated by respective extensive numerical experiments. Managerial insights are also derived for relevant stakeholders.