Abstract: Dynamic flight scheduling means updating scheduling scheme timely, when the queue of the flight to be scheduled changes. This requires that scheduling algorithm should obtain stable sort results. An improved genetic algorithm based on MPS restriction is proposed to solve the problem that the results of GA aren’t stable. It can avoid effectively the situation of premature convergence and improve efficiency of search. The results show that the scheduling algorithm can get stably the optimal solution.

Key words: genetic algorithm, dynamic flight landing scheduling, maximum position shift, penalty function, global optimization solution