Abstract:To address declining service efficiency, intensified pedestrian-vehicle conflicts, and spatial resource mismatch in ride-hailing pick-up zones of large integrated transportation hubs during peak periods, this study develops a hybrid simulation framework combining Agent-Based Modeling (ABM) and Discrete Event Simulation (DES) to analyze collaborative “pedestrian-vehicle-facility” scheduling. Four scenarios are designed: routine peak, holiday mass passenger flow, extreme peak, and emergency events. Through vehicle-flow gradient experiments, stress tests, and multi-dimensional optimization strategy evaluation, the degradation patterns of service levels and critical carrying capacity thresholds are quantified. Results show that zonal operation meets evacuation demand under routine and holiday peak scenarios; under extreme peak conditions, average passenger waiting time exceeds the 15-minute warning threshold and enters a nonlinear deterioration phase. Passenger flow diversion, improved connection efficiency, and “vehicle-seeking-passenger” optimization significantly enhance service levels. This research provides quantitative decision support for tiered response planning, dynamic control optimization, and emergency management in hub ride-hailing connection systems.