Yao, Xuejian (Southeast University), Zhang, Jian (Southeast University), Wang, Bo (Southeast University), Qian, Yu (Southeast University), Wang, Maoze (Southeast University)

Research on Dynamic Resource Supply Strategy of Hard Shoulder of Expressway Based on Reinforcement Learning

Scheduled for presentation during the Invited Session "S17c-Synthetic-Data-Aided Safety-Critical Scenario Understanding in ITS" (TH-LA-T17), Thursday, November 20, 2025, 16:00−16:20, Southport 2

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 18, 2025

Abstract

Some aging expressways with limited lanes are unable to provide sufficient service resources when encountering sudden traffic surges caused by incidents or emergencies. Effectively utilizing existing hard shoulder emergency lanes for dynamic resource supply has become a new objective in the context of intelligent capacity expansion. A dynamic hard shoulder resource supply method is proposed. Firstly, a scenario model for basic expressway mainline lanes and hard shoulders is established using the Cell Transmission Model (CTM). Secondly, a fully centralized multi-agent system for hard shoulder control is constructed, and a reinforcement learning-based strategy is developed for dynamic resource supply. Finally, a series of simulation experiments are conducted to compare the proposed strategy with the baseline strategies of full opening and full closure. The results demonstrate that the proposed method reduces the average vehicle cell travel time by 51.7% and the congestion dissipation time by 39.6%, effectively supporting dynamic resource supply.