Research on Interval Estimation of Trip Fuel Consumption Based on Irregularly Distributed Samples

Interval estimation of aircraft fuel consumption is an important basis for airlines to make flight plans and to save energy and reduce emissions. Aiming at the problem that the fuel consumption samples under the influence of multiple factors presented irregular distribution, which would reduce the rationality and accuracy of fuel consumption estimation, an interval estimation method of trip fuel consumption based on the combination of support vector quantile regression and Bootstrap (SVQR-Bootstrap) was proposed. The method based on quantile regression residual error minimization of the absolute value of asymmetric form thoughts to obtain the conditional quantile of trip fuel consumption and on this basis, through the Bootstrap estimate mean and standard deviation of trip fuel consumption population, and then, according to fixed aircraft type and segment, the population of fuel consumption meets the normal distribution, building estimated interval of trip fuel consumption under a constant confidence level. According to simulation comparisons, the SVQR-Bootstrap method can effectively reduce the length of the estimation range, pointing to the clear, which is better than the interval estimation method which requires assumed data distribution.