A Recommendation for Flight Extra Fuel Based on Random Forest

It is critical for airlines to develop an effective fuel-saving strategy because the fuel cost accounts for a large proportion of the operating costs in air transport industry. This paper aims to optimize the flight extra fuel predicted by dispatchers via data mining techniques on the consideration of both economy and safety. We first establish a random forest (RF) model from history data to predict the consumption of trip fuel. Then to ensure safety index, we add a margin to the output of the model as the modified prediction value. Finally, the experimental results on real flight data show that the proposed method can make a more accurate prediction for the extra fuel adding amount under the condition of ensuring safety. It indicates that the proposed method is beneficial for airlines in economic cost saving.