A Heuristic Approach for Solving Flight Retiming Problems

Flight delays are a problem faced by airlines every day. A flight delay can induce delay propagations to other flights, and it results in the increasing airline operating costs. To deal with this problem, airlines should plan flight schedules that are not sensitive to be delayed. Revising flight schedules is one way to allocate slack to flight connections which are potential to induce flight delay. This paper investigates a problem to revise departure time of flights. We derive an optimization model for retiming departure time of flights by defining a nonlinear function to measure how close the new slack to the estimated slack. We limit the change of the departure time in short interval of time to preserve passenger projections in flight schedule planning stage. To reduce the complexity, the model is solved by defining a set of flight copies for each flight, then genetic algorithm is applied to select the best flight copy. We test the model to the flight schedules of an airline in Indonesia. The computational results show that our approach produces new schedules which almost all estimated slack distribute to schedule connections. Moreover, the new schedules able to decrease the number of flight delay and its delay duration.