Dynamic Traffic Assignment Models for System Optimal Future Mobility Analysis

Abstract

System optimum dynamic traffic assignment (SODTA) models predict a time-dependent traffic state with optimal network performance, providing a benchmark for controlling and managing dynamic traffic networks. This thesis explores the applications of these models for congestion mitigation based on a novel optimisation framework for system-level future mobility analysis.

This thesis has three aims: (1) to provide a mathematical foundation for developing a framework for network-level analysis of traffic flow, (2) to explore the usefulness of the proposed model for various network-level design problems with advanced congestion mitigation strategies, and (3) to explore the practicality of the proposed model for futuristic transport scenarios in an automation heavy network. These aims are achieved and presented in the three core chapters of this thesis. The first core chapter develops the base model of SODTA embedding the link transmission model (LTM) for dynamic network loading and traffic flow propagation and implements it on single-OD and multi-OD networks.

The second core chapter explores three strategies for congestion mitigation with system-level mobility analysis based on further development of the base model. The three strategies involves a classical example of network design problem with potential capacity enhancements, a departure time incentive scheme to encourage commuters to shift their departure times to maintain an optimal system performance and a shared mobility service to cater to the travel demand of a network where commuters are incentivised to share their rides to reduce overall congestion in the network.

Finally, the third core chapter develops the base model even further to analyse a network which includes both legacy vehicles (LVs) and vehicles with automation features such as cooperative adaptive cruise control, speed harmonisation and cooperative merging. Here, an integrated mixed-integer programming framework is proposed for optimal exclusive lane design for these automated vehicles (AVs) on a freeway network which accounts for commuters' demand split among AVs and LVs via a logit model incorporating class-based utilities.

Overall, this thesis exploits the potentialities of SODTA models to evaluate futuristic transport scenarios. The recent technological advancements in transport industry indicate an exponential rise in cooperation and coordination between transportation system and its stakeholders rendering these models essential tools for future mobility analysis.