| Author: | Huang, Yunping |
| Title: | Dynamic traffic assignment models for urban network : a macroscopic fundamental diagram approach |
| Advisors: | Hsu, Shu-chien (CEE) Sumalee, Agachai (CEE) |
| Degree: | Ph.D. |
| Year: | 2024 |
| Subject: | Traffic flow Urban transportation Hong Kong Polytechnic University -- Dissertations |
| Department: | Department of Civil and Environmental Engineering |
| Pages: | xx, 209 pages : color illustrations |
| Language: | English |
| Abstract: | This dissertation involves the development of three key components of dynamic traffic assignment models, i.e. network loading model, dynamic traffic assignment principle, and network optimization with available/unavailable network MFD information. The imbalance between network capacity and travel demand is the core reason for traffic congestion. Increasing road supply becomes unsustainable due to the limited land space in crowded urban cities. Route guidance, traffic control, and dynamic pricing are adopted by traffic authorities and managers to adjust traffic flows to optimize travel demand distribution and alleviate traffic congestion. Dynamic traffic assignment (DTA) provides a benchmark for evaluating these congestion management measures. Dynamic network loading (DNL) captures traffic flow propagation and describes how travelers’ route and departure time choices affect travelers’ travel costs. DTA highly relies on the loading models, while most loading models are based on link-level. Due to the large spatial dimension of urban networks, detailed link-level models are computationally intensive and intractable in the real world. Macroscopic fundamental diagram (MFD), describing the aggregate relationship between the network flow accumulation and the space-mean traffic variables at the network level, initiates a promising solution to the challenge of spatial dimensionality. Therefore, this dissertation proposes an MFD-based network loading model and considers the travelers’ behavior assumptions, i.e., dynamic user equilibrium (DUE). Three MFD models, the accumulation-based model, the trip-based model, and the time delay model, were proposed in the literature to capture traffic flow propagation. However, no consensus has been reached on their computational efficiency and which model should be chosen under certain traffic conditions and demand scenarios. Study 1 revisits these models regarding two important theoretical properties regarding flow propagation in the DTA, i.e., the first-in-first-out (FIFO) principle and causality. Numerical studies validate that the time delay model provides a good approximation in both free-flow and saturation scenarios without violating strict causality. Thus, the time delay model is a promising alternative for dynamic network loading in large-scale network applications. Using the time delay model to describe the multi-region MFD dynamics, Study 2 presents an optimal control framework to model route choice behavior with departure time choice, i.e., DUE, for general urban networks. It is assumed that the drivers would choose their departure times and routes to minimize their generalized travel costs. Difficulties raised in handling the dynamic state-dependent nonlinear travel time functions, state, and inflow constraints are addressed without model linearization nor enforcing constant delay assumption as conventionally done in the literature. The additional cost induced by inflow capacity and accumulation constraints can capture the hypercongestion represented by the downward-sloping part of the MFD without actually activating traffic congestion. Numerical examples illustrate the characteristics of DUE and their corresponding dynamic external costs induced by constraints. For ride-sourcing systems, the imbalance between vehicle supply and demand is a long-standing challenge, leading to losses of orders and thus additional travel burdens to the network. Contrary to DTA optimizing the travel demand distribution, ride-sourcing systems optimize the supply distribution by dispatching and relocating vehicles to satisfy passenger demand. Nevertheless, they face similar computational challenges. Study 3 and Study 4 investigate fleet optimization with available and unavailable MFD information, respectively. Optimal relocation of idle vehicles to high-demand regions can enhance its efficiency, improving the quality of service and reducing the overall congestion of the whole network in the long run. Enforcing vehicle relocating with link-node representation or grid-based representation is hard to capture the interrelated dynamics with private vehicles in addition to its computationally intensive nature. The macroscopic fundamental diagram (MFD) provides a powerful tool to model those dynamics while individual vehicle details may be absent with regional-level representation. Therefore, Study 3 proposes a bi-level rebalancing scheme to maximize the served orders in the system. For the upper level (network level), the interrelated dynamics of private vehicles and taxis are modeled based on the MFD. Then a stochastic programming problem is formulated and solved using the approximate dynamic programming (ADP) algorithm to determine the number of desired vehicles in each subregion and cross-border. For the lower level (i.e., vehicle level), a Voronoi-based distributed coverage control algorithm is implemented by each vehicle to obtain position guidance efficiently. The bi-level framework is evaluated on a simulator modeling the real road network of Shenzhen, China. Simulation results demonstrate that, compared to other policies, the proposed bi-level approach can serve more requests with less waiting time while reducing the overall congestion and externalities of the network. However, ride-sourcing service operators may not have access to the information of the network MFD to capture dynamic travel time. However, uncertain demand and dynamic travel time subject to traffic congestion can significantly affect the optimal solutions. To meet these challenges, we first estimate the network-level journey time with functional data analysis capturing the dynamics and stochasticity of travel time. We then propose a multi-stage decision model to address the matching and decision of a fleet of vehicles for a centralized platform. Furthermore, we formulate the problem as a stochastic programming model to account for spatial-temporal uncertainties in customer demand. And we developed an Approximate Dynamic Programming (ADP) based approach to efficiently solve the multi-stage decisions. To evaluate the effectiveness of our algorithm, we utilize a designed simulator based on NYC yellow taxi data and the Manhattan road network. Numerical studies demonstrate that the use of time-dependent travel time data is beneficial in terms of improving system profit compared to using the mean historical travel times. ADP significantly enhances total system profit compared to several popular decision practices. In conclusion, this thesis contributes to the literature on macroscopic fundamental diagram-based dynamic traffic assignment and its applications in ride-sourcing systems. |
| Rights: | All rights reserved |
| Access: | open access |
Copyright Undertaking
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
- I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.
Please use this identifier to cite or link to this item:
https://theses.lib.polyu.edu.hk/handle/200/12798