https://hal-enac.archives-ouvertes.fr/hal-01136266Deschinkel, KarineKarineDeschinkelFEMTO-ST - Franche-Comté Électronique Mécanique, Thermique et Optique - Sciences et Technologies (UMR 6174) - UTBM - Université de Technologie de Belfort-Montbeliard - ENSMM - Ecole Nationale Supérieure de Mécanique et des Microtechniques - CNRS - Centre National de la Recherche Scientifique - UFC - Université de Franche-Comté - UBFC - Université Bourgogne Franche-Comté [COMUE]Farges, Jean-LoupJean-LoupFargesONERA - The French Aerospace Lab [Toulouse] - ONERADelahaye, DanielDanielDelahayeLOG - Laboratoire d'Optimisation Globale - ENAC - Ecole Nationale de l'Aviation CivileDGAC - Direction Générale de l'Aviation Civile Optimizing and assigning price levels for air traffic managementHAL CCSD2002Gradient algorithmSimulated annealingAir trafficPricing policies[MATH.MATH-OC] Mathematics [math]/Optimization and Control [math.OC]Porte, Laurence2015-03-26 19:34:302022-04-21 09:50:022015-03-26 19:34:30enJournal articles10.1016/S1366-5545(02)00007-81Pricing policies could encourage airline companies to modify departure times and routes of their flights in order to reach a route-slot allocation target that minimizes the en-route congestion. The problem of restricting the number of price levels and the assignment of one price level to each sector at each time period is studied using a simulation based on a Logit discrete choice model. In this model, an option is defined as the combination of a departure time and a route. For each flight, a utility is associated to each option and takes into account the flying cost, the cost of ground delay and the prices of crossed sectors. The optimization of the pricing policy considers average flows and minimizes the quadratic difference between desired and expected flows on each option. A heuristic algorithm that involves successive iterations of simulated annealing and gradient methods performs this optimization. The simulated annealing assigns a price level to each sector at each time period, and the gradient algorithm computes new values of price levels. The test of the method on constructed examples indicates that the use of only four price levels does not significantly deteriorate the performance of the system with respect to the use of independent prices.