Dynamic Hard-Shoulder Running Control for Expressways Based on an SIR-T Model: Advancing Sustainable Traffic Management

With the continuing growth of expressway travel demand, recurrent congestion frequently occurs on selected expressway sections during peak periods. Conventional fixed-threshold strategies for hard-shoulder running have difficulty maintaining a dynamic balance between congestion mitigation, safety-related operational considerations, and lane-resource utilization. This study proposes an SIR-T (Traffic) state-evolution model for hard-shoulder opening and closing decisions. Traffic operation is represented by three normalized states, namely, free flow or congestion susceptibility (Susceptible, S), congestion formation and propagation (Infected, I), and congestion recovery (Recovered, R). The state-transition equations describe the effective congestion propagation rate, recovery transition rate, and time-dependent return rate. Literature-informed and section-specific baseline parameters are adopted, while hourly electronic toll collection (ETC) gantry traffic volumes from the Zhejiang G-YZ Expressway are used as the time-varying demand input for numerical integration. Dynamic opening and closing conditions are then derived and evaluated in SUMO by comparing the proposed control strategy with static opening and no-opening strategies. The results show that the proposed dynamic control strategy can adaptively adjust the hard-shoulder opening period in response to real-time traffic conditions, reduce unnecessary occupation of shoulder resources, and maintain traffic service performance close to that of static opening while producing fewer lane changes; compared with no opening, it also reduces average travel time and increases total arrivals.