Abstract
Traffic congestion is known to have many negative impacts both for travelers and society as a whole (e.g., emissions, noise). This dissertation focuses on the model development and analysis of vehicle and trip flow dynamics. Several transportation systems are considered, from a single bottleneck to a network region, in order to propose effective management strategies to reduce traffic congestion. Since different transportation systems are considered, different models and space paradigms are needed to model each system’s dynamics. At the local level, it is reasonable to study the traffic flow in the physical road and how it evolves spatially and temporally with traditional traffic flow theory models. In contrast, when the network is highly complex (e.g., corridors with many on- and off-ramps or urban cores), it is reasonable to disregard the detailed vehicle-network interaction at the distinct physical locations and model traffic dynamics with more aggregated models in the relative space. The relative space differs from the traditional absolute space because the former is defined relative to individual trips’ destinations, and the trajectories of different trips in the network but with different origins and destinations can be studied together in the same (relative) space-time domain.
When queues start to build up at local freeway bottlenecks, the capacity drop (CD) phenomenon reduces the maximum flow that the bottleneck can accommodate. First, the heterogeneity of vehicles’ driving parameters is studied for such local transportation systems. On the management side, some researchers propose to use Variable Speed Limit control (VSL) to prevent the CD. However, the necessary acceleration length between the end of the control application area and the bottleneck has not been analyzed thoroughly. By developing an effective open-loop control, I present an analytical formulation to determine the minimum acceleration stretch required to prevent the CD.
The so-called “bathtub model” captures the inflow, outflow, and the instantaneous number of active vehicles in the network by assuming: (i) that the network can be treated as an undifferentiated unit where vehicles travel in a relative space towards their own destination, and (ii) that there is a network-level speed-density relationship for all vehicle trips. This bathtub model and the relative space perspective are attractive to model complex network systems. However, most studies disregard the role that demand plays in such a system. The trip distance distribution (TDD) is part of the bathtub model and is not well understood. In this dissertation, empirical data from Chicago is used to show that most of the existing assumptions on TDD do not hold. Further, I propose to study the trip flow dynamics by developing a probabilistic agent-based bathtub model, i.e., a microscopic simulation model in the relative space, to track the completion rate of trips and trip duration of individuals, given any TDD. Then, the bathtub model is used to study the corridor level dynamics and propose a dynamic distance-based high occupancy toll lane pricing scheme. Further, several fleet sizing strategies of shared mobility systems are explored. To do so, a compartmental model for trip flow dynamics is proposed, where trips can be waiting for a shared vehicle to pick them up (point queue model) or traveling in the network (bathtub model).
In summary, this dissertation presents a comprehensive review of two modeling paradigms where space is treated differently. It explores particular management strategies leveraging the most suitable modeling paradigm to improve traffic congestion. In this dissertation, the traffic dynamics of different transportation systems are studied by integrating first-principles analysis, data-driven methods, and simulation-based studies. This dissertation lays a good foundation for future studies on emerging mobility systems.
