Using Genetic Algorithms to Evaluate Aircraft Ground Holding Policy in Real Time

Abstract

Ground holding policy (GHP) offers a potential solution to alleviate increasing congestion problems in the air transportation network. The policy proposes to hold an aircraft before takeoff so that when it arrives at its destination, it need not wait in the air before landing. This is motivated by the fact that airborne delays are costlier than ground delays. Many attempts have been made to solve the static GHP problem. However, aspects such as flight cancellations, transfers at hub airports, en route speeding, weather changes, and the interdependent nature of arrival and departure capacities suggest that updating ground holding decisions as time progresses is important. In this paper the dynamic instantaneous GHP problem, considering some of the above aspects, is formulated as an integer programming model. The presence of binary integer variables and discrete variables make the GHP problem complex and computationally difficult to solve using traditional algorithms. As an alternate approach, the use of genetic algorithms is demonstrated and analyzed in this paper. The observed program run time is reasonably small, indicating that genetic algorithms can be used for implementing GHP in real time for dynamic operating environments.