My research interests focus on optimization algorithms and Machine Learning techniques applied to Air Traffic Management (ATM) problems.
Many Air Traffic Management problems can be directly modeled as constrained optimization problems, and/or require the use of optimization techniques for solving them. For example, when solving conflicts between aircraft trajectories, one needs to minimize the trajectory deviations while separating the aircraft. Other ATM problems, such as the prediction of air traffic controller workload or the prediction of aircraft trajectories, can be addressed by Machine Learning techniques.
Practically, the choice of the optimization method depends on the problem being addressed and the chosen model. Different algorithms might be used, depending on the function to optimize (analytical expression, existence of derivatives, or "black box"), the type of variables (discrete, real, or stochastic variables), and the size and difficulty of the problem.
My most recent research, in collaboration with other researchers at ENAC/MAIAA and IRIT/APO, deals with the cooperative parallelism of metaheuristics and interval methods. The aim to find and prove the global optima of high-dimension multimodal functions that are difficult to optimize. See the optimization page of the IRIT/APO web site for more details on these two types of methods: metaheuristics and interval branch & bound algorithms.
Another part of my research activity consists in applying Machine Learning techniques to problems such as the prediction of the air traffic controllers' workload, or the prediction of aircraft climbs.
I am currently working on the following ATM subjects:
For memory, here is a list of research activities in which I was involed in the past years:
Last Update: |