A robust optimization approach for dynamic airspace configuration

Abstract

Many factors are contributing to raising challenges in Air Traffic Management operations, from increasingly adverse weather conditions to emerging usages of airspace. In this context, efficiently managing limited airspace capacity while accounting for traffic demand uncertainty has become critical. Dynamic Airspace Configuration provides a framework to maximize efficiency by adapting airspace capacity to varying spatial and temporal demand patterns, thereby minimizing traffic overflow and reducing regulations and delays. Given a pre-determined set of configurations, we aim to determine an optimal and robust configuration plan that effectively absorbs air traffic under demand uncertainty. We propose two solution approaches: an integer linear programming model and a more computationally efficient graph-based formulation using a constrained shortest path algorithm. We extend the formulations to account for uncertainty and provide optimal configuration plans that are robust against possible traffic demand increase, with different levels of protection. We evaluate our robust approach to dynamic airspace configuration on Madrid ACC, considering available configurations and traffic data from August 2024. Our computational study explores trade-offs between minimizing traffic overflow and robustness, demonstrating that even moderate levels of conservatism can significantly impact traffic excess and, consequently, delays. These findings underscore the importance of computing optimal robust solutions.