Parishad, Nasser (the University of Queensland), Yildirimoglu, Mehmet (University of Queensland), Hickman, Mark (the University of Queensland)

Distance-Based Pricing in Multi-Modal Urban Networks with Deep Reinforcement Learning

Scheduled for presentation during the Regular Session "S09b-Optimization for Multimodal and On-Demand Urban Mobility Systems" (WE-EA-T9), Wednesday, November 19, 2025, 14:50−14:50, Coolangata 3

2025 IEEE 28th International Conference on Intelligent Transportation Systems (ITSC), November 18-21, 2025, Gold Coast, Australia

This information is tentative and subject to change. Compiled on October 19, 2025

Abstract

Among various congestion mitigation strategies, congestion pricing emerges as particularly effective in reducing peak-time congestion. However, developing a dynamic, real-time, equitable pricing mechanism that proactively generates toll profiles while considering demand elasticity and travellers' heterogeneity is a challenging task. This study introduces a data-driven, distance-based pricing framework using reinforcement learning to optimise multi-modal traffic conditions in congested urban areas. A trip-based three-dimensional macroscopic fundamental diagram (3D-MFD) simulation has been developed, capable of capturing individual mode-choice decisions. Traveller heterogeneity is addressed through variations in origin and destination, trip length, departure time, and value of time (VoT). To establish a tolling strategy that maximises network outflow (minimises total travel time), a Deep Q-Network (DQN) agent has been introduced. Without any prior knowledge of network parameters, the agent successfully alleviates traffic congestion and reduced total travel time by approximately 60%.