Abstract
The air traffic demand is rapidly growing in recent years on the background of economic growth of Asia-Pacific area and popularization of Low-Cost Carriers (LCCs). Every aircraft is planned to fly along the flight plan. However, most of all aircraft are not able to follow the scheduled plans due to various factors. One of the severe factors is the over-capacity of the arrival airport. When the number of aircraft exceeds the capacity of the airport, the air traffic controller does not allow airplanes to land. In this case, aircraft have to take a detour, so-called "vectoring," or keep making circles on the particular points, so-called "holding" to adjust the arrival time. These time adjustments affect not only the delay of arrival time but also an increase in fuel consumption. If there is an efficient way to prepare a schedule with less time adjustment beforehand, it will be useful for future air traffic control. The objective of this study is to make efficient flight schedules with less congestion and enough resilience against traffic problems. In this study, we conduct a preliminary multi-objective optimization for aircraft by using NSGA-II. The objective functions are (1) minimization of averaged arrival-delay, (2) minimization of the ratio of the number of delayed aircrafts, and (3) minimization of the mean fuel consumption. The design variables are the departure-time offset of each domestic aircraft for Tokyo International Airport. We adopt a cellular automaton for simulating air traffic. As the result of optimization, a lot of optimal solutions, so called non-dominated solutions, are obtained. These solutions can decrease not only the arrival-delay time but also fuel consumption. The results indicate that the slight takeoff-time differences as the result of the adjustment of the spot assigns and taxing routes effect profoundly on the air traffic flow.
