Application of artificial fish school and K-means clustering algorithms for stochastic GHP

In order to make full use of airport capacity and eliminate existing human errors, typical capacity scenarios are produced, based on artificial fish school and K-means clustering algorithms. Then nominal capacity scenarios tree is constructed, which can be used in stochastic GHP model. Compared to case of no-GHP, the delay cost in static and dynamic models is reduced by 37.2% and 57.2% respectively. It is concluded that the mixed algorithm is feasible and the nominal capacity scenarios tree is practical.