A Clustering-based Single-parent Genetic Algorithm for Airport Logistics Vehicle Scheduling

The effective scheduling of airport ground logistics vehicles is crucial for improving the quality and efficiency of airport logistics services and enhancing the coordination of airport operations. This study focuses on the airport logistics vehicle scheduling problem (ALVSP), which aims to optimize logistics support operations (cargo and baggage support tasks) within a period of time. By considering each support task as a node, we construct a complete directed graph to represent different sequences of tasks. Then we model ALVSP as a single-depot asymmetric multiple traveling salesman problem (MTSP) and propose a clustering-based single-parent genetic algorithm (CSPGA) to address it. Specifically, a disruption and reclustering mechanism (DRM) is integrated into the single-parent genetic algorithm (SPGA) to achieve dynamic clustering, an elite selection strategy is employed to retain the best solutions discovered during iterations, and an adaptive mutation operator selection mechanism (AMOSM) is leveraged to dynamically select the most effective mutation operator from the candidate operator pool according to their historical performance. Furthermore, a local search approach is leveraged to further optimize the solutions after the genetic operations. Finally, we evaluate the performance of CSPGA on different datasets, and the results demonstrate that CSPGA outperforms the comparison algorithms on most instances.