Multi-objective optimization of flight-gate assignment based on improved genetic algorithm

In order to improve the resource utilization efficiency and passenger transfer experience of modern airports, the multi-objective flight-gate assignment problem was studied. Considering the constraints of flight type, aircraft body type and transition time interval, a multi-objective nonlinear 0-1 integer planning model of the flight-gate assignment was established by taking the maximum number of flights allocated at a fixed gate, the minimum number of used fixed gates and the minimum passenger transfer tension as the objective functions. Then a genetic algorithm based on the improved gene coding was designed to improve the solving efficiency of the model. The gene individual adopts two-stage integer coding, and the mapping process from the gene coding method to a feasible solution was designed. Meanwhile, it was theoretically proved that the gene coding method could be mapped to the optimal solution. Different crossover operators and mutation operators were designed for the two stages of gene coding to avoid infeasible individuals. In order to verify the effectiveness of the algorithm, based on the actual operation data of a large-scale airport, the improved genetic algorithm and MATLAB built-in genetic algorithm were compared. Calculation result shows that with the improved genetic algorithm, the number of flights assigned to the fixed gates increases by 5%, the total transfer tension of passenger decreases by 3%, the average transfer tension of passenger decreases by 32%, the number of used fixed gates stays the same, the passengers assigned to the fixed gates increase by 20%, and the running time of the algorithm reduces by 8%, which shows that the improved genetic algorithm has better performance and can improve the gate utilization efficiency and passenger transfer comfort. In the optimization process of the improved genetic algorithm, the number objectives of flights and gates reach the best in 130 iterations, the transfer tension basically converges after 400 iterations, and the flight schedule generated by the optimal solution is reasonable, which indicates that the algorithm has fast iterative convergence speed and reasonable optimization result. © 2020, Editorial Department of Journal of Traffic and Transportation Engineering. All right reserved.