Plans de vols Pareto-optimaux : coût du carburant vs durée du vol

点击翻译按钮获取中文摘要

Flight plan optimization has to make a trade-off between fuel cost and flight time. For example, if a flight is delayed by a few minutes, as is often the case, one has to determine how much fuel should be used in order to increase the speed of the aircraft and, thus, reduce the costs resulting from a delayed arrival. This trade-off is currently obtained by minimizing the fuel cost plus the cost index times the flight time. This method has two drawbacks. First, the cost index has to be determined by the planner, and often many attempts are necessary before a satisfactory value is found. Second, fuel cost as a function of flight time is not convex for all duration values. Flight level changes, tariffed zones, prohibited airspace or adverse wind conditions can lead to time intervals where the function is not convex. As a result, no value of the cost index makes it possible to obtain a flight time in such an interval. Yet it is possible to have a flight plan corresponding to this flight time.This project develops a new formulation of this problem and a new solving method with which optimal flight plans can be obtained for all durations in a few seconds. Two systems are developed: one for flight planning, where the quantity of fuel to be taken on board has to be determined, the other one for updating the flight plan, where the quantity of fuel is fixed, and one has to go through some points within a time window with minimum fuel supply.In addition to including the fundamental research objectives on flight plan optimization, this project aims at bringing the new developed systems to a high level of maturity so that they quickly become usable by industrial end users.