Soutenance de thèse : Mme Niloofar SHAKOORI

Optimal Dynamic Routing for Urban Networks: a Mathematical Programming Approach with Complete Integration of Traffic Flow Features

Doctorante : Mme Niloofar SHAKOORI

Laboratoire : LICIT
Ecole doctorale : ED162 MEGA de Lyon (Mécanique, Energétique, Génie civil, Acoustique)
 

Urban transportation networks face increasing challenges from congestion and environmental impacts, necessitating a balance between system efficiency and environmental sustainability. Optimal routing strategies are essential but traditional Dynamic Traffic Assignment (DTA) models often fall short due to their reliance on triangular fundamental diagrams, which may not accurately represent urban traffic complexities. Moreover, existing frameworks generally prioritize travel time over environmental objectives like emissions, underscoring the need for a refined approach that integrates realistic traffic features with optimization frameworks addressing both efficiency and environmental impact.

This thesis advances System-Optimal Dynamic Traffic Assignment (SO-DTA) literature by introducing a Link Macroscopic Fundamental Diagram (MFD)-based traffic model to better represent urban traffic dynamics. This model is integrated within a Mathematical Programming (MMP) context using piecewise linear (PWL) functions, designed to optimize both total system travel time (TSTT) and total system emissions (TSE) in general networks. Furthermore, the framework rigorously incorporates the Vehicle Holding (VH) problem and the First-In-First-Out (FIFO) principles simultaneously. While literature often addresses VH and FIFO separately due to their complexities, including both enhances the operational effectiveness of DTA models.

Key contributions of this thesis include: (1) a novel link MFD-based traffic model that captures urban traffic dynamics more accurately, (2) integration within an MMP framework employing PWL approximations, (3) dual-objective optimization for TSTT and TSE, enhancing applicability in urban contexts, (4) refined NVH formulation within the PWL formulation of the link MFD-based traffic model, and (5) explicit integration of the FIFO principle to model operational effectiveness. Additionally, solution strategies, including rolling horizon techniques and path selection strategies, are proposed to address computational challenges and improve the scalability and performance of the proposed framework.