An improved optimization algorithm for solving arrival aircraft scheduling problem in the Terminal Maneuvering Area

Abstract

The Terminal Maneuvering Area (TMA) is one of the most complex and congested airspace segments, where tools like the Arrival Manager (AMAN) are used to manage inbound traffic and provide accurate, efficient scheduling for each flight. The associated optimization problem is NP-hard, requiring advanced algorithms to meet performance demands in both computational time and solution quality. Heuristic algorithms, such as Simulated Annealing (SA), are known for their ability to provide fast, near-optimal solutions in large and complex state spaces. In our previous work, simulation-based optimization using SA was employed, where information of all flights was integrated into each simulation, resulting in a computationally intensive evaluation process. In this study, we propose a more efficient method by leveraging the inherent safety dependencies between neighboring flights in the operation.By focusing on the performance of individual flights and their immediate impact on adjacent flights, the optimization process becomes more targeted, eliminating the need to integrate all flight data at once. This improves both efficiency and flexibility. To demonstrate the advantages of a selective structure in Simulated Annealing, we introduce Selective Simulated Annealing (SSA) and compare it to the Standard Simulated Annealing algorithm (OSA), highlighting their distinct features. A case study at Paris-Charles de Gaulle (CDG) Airport is used to analyze the performance of both algorithms. Key parameter adjustments are examined to gain insights into their optimization behaviors. The comparison reveals that SSA significantly outperforms OSA, delivering faster computation and reducing delays by 50%.