Enhancing Best-First Search Algorithms for Aircraft Ground Trajectory Optimization With Non-Additive Cost Functions
Adrien Durand, Georges Ghazi, Ruxandra Mihaela Botez · 2024
This paper presents a study conducted at the Laboratory of Applied Research in Active Controls, Avionics, and AeroServoElasticity (LARCASE) for optimizing aircraft ground trajectories at airports using non-additive cost functions. For this purpose, the airport was modeled as an undirected graph composed of nodes and edges, with edge properties varying depending on the aircraft path. Consequently, the value of the cost function can not be determined by summing the cost of individual edges, but it also depends on the overall aircraft path. To address this challenge, a modified bi-directional A* algorithm was developed to find the optimal path between two nodes in the graph that minimizes fuel consumption or taxiing time, while satisfying geometrical constraints. The cost function used in the optimization algorithm was defined as non-additive, making the proposed optimization algorithm suitable for solving more complex problems where traditional Best-First Search (BFS) methods, such as Dijkstra or A* algorithms, fail to provide optimal solutions (considering non-additive cost functions). The proposed algorithm was tested at Montreal (CYUL) Airport using a fuel consumption model of the Cessna Citation X business jet aircraft derived from a Level-D Research Aircraft Flight Simulator (RAFS). The results showed that it was possible to reduce the aircraft fuel consumption by an average of 3.86% and the taxiing time by an average of 4.10%.