Simulation-Based Two-Stage Scheduling Optimization Method for Carrier-Based Aircraft Launch and Departure Operations.

The scheduling of carrier-based aircraft departure operations is subject to stringent temporal, spatial, and resource constraints. Conventional approaches struggle to yield exact solutions or provide a comprehensive mathematical description of this complex, dynamic process. This study proposes a simulation-based optimization method, establishing a high-fidelity simulation model for aircraft departure scheduling. To address the coupled challenges of path planning under spatial constraints and station matching/sequencing under operational constraints, we developed (1) a deep reinforcement learning (DRL)-based path planning algorithm (AAE-SAC), and (2) an enhanced particle swarm optimization (PSO) algorithm (LTA-HPSO). This integrated two-stage framework, termed LTA-HPSO + AAE-SAC, facilitates efficient, collision-free departure scheduling optimization. Simulation experiments across varying sortie scales were conducted to validate the framework's effectiveness and robustness. Notably, for a complex scenario involving 24 aircraft with diverse priorities and stringent spatial constraints, LTA-HPSO + AAE-SAC achieved an average solution time of 185.19 s, reducing scheduling time by 26.18% and 49.54% compared to benchmark algorithms (PSO + Heuristic and PSO + SAC, respectively). The proposed LTA-HPSO + AAE-SAC framework significantly enhances the quality and robustness of carrier-based aircraft departure scheduling.