Integrating agent-based simulation and optimization for dynamic resource scheduling at airport security checkpoints

초록

The rapid growth of global air travel has intensified operational pressures on airport security checkpoints, where passenger throughput and service reliability must be jointly maintained. While existing studies often assume fixed or exogenous passenger-arrival profiles, such static approximations fail to capture the temporal variability observed in real operations. Building upon recent advances in simulation-based modeling, this study proposes a data-driven simulation-optimization framework for the dynamic scheduling of airport security screening checkpoints (SSCPs). The framework integrates a stochastic optimization model that generates delay-aware laneopening schedules, a mixed-integer linear programming (MILP) projection that ensures workforce-feasible implementation under staffing constraints, and an agent-based simulation (ABS) that evaluates operational performance using LiDAR-derived passenger-arrival data and empirically calibrated service times. The LiDARbased arrival reconstruction enhances temporal accuracy, enabling the model to reflect short-term fluctuations in passenger inflows and adapt resource allocation accordingly. Empirical analysis using real airport data shows that the proposed framework accurately reproduces observed passenger flow patterns and reduces passenger waiting times by 6.4% under normal demand conditions and by 10.3% during holiday peak periods, without increasing total staffing capacity. These findings demonstrate that intelligent temporal redistribution of existing screening resources can significantly enhance operational efficiency and passenger service quality.

키워드