Nowadays, constellations of satellites have to deal with heterogeneous and complex observation requests, such as one-shot, video, stereoscopic, and periodic requests. In this paper, we consider the problem of scheduling these requests in order to maximize a measure of global utility. To solve this problem, we propose two Large Neighborhood Search algorithms that exploit problem decompositions. These algorithms explore large neighborhoods respectively based on heuristic search and Constraint Programming. The experiments performed on instances generated from real constellation features and weather data show that the approaches improve the state of the art.
Lieu : grande salle de réunions dans la coursive du bâtiment R de l'ONERA à Toulouse.
Lien visio : https://rdv.onera.fr/seminaireDTIS