基于改进GA的神经网络航班保障时间估计

准确估计航班保障服务时间可以极大提高地面航班保障服务效率。采用主成分分析(PCA)方法降低变量间的相关性,考虑到BP神经网络的网络结构难以确定,且网络初始权重、阈值随机,提出改进的遗传算法来优化BP神经网络的结构,初始权重、阈值,建立自适应多层遗传算法(AMGA)的BP神经网络航班保障服务时间估计模型。为验证所提AMGA-BP算法的性能,以国内某枢纽机场航班保障服务时间作为研究对象,与传统的GA-BP、BP两种算法做对比实验,进行航班保障服务时间估计,实验结果表明,AMGA-BP算法比BP算法和GA-BP算法精确度更高。

Estimation of flight support time based on improved GA neural network

Accurate estimation of flight support service time can greatly improve the efficiency of ground flight support services.The method of principal component analysis(PCA)was used to reduce the correlation between variables.Considering that the network structure of the BP neural network is difficult to determine,and the initial weight and threshold are random,the improved genetic algorithm was proposed to optimize the network structure.The weight and thresholds were initialized simultaneously,and the adaptive multi-layer genetic algorithm(AMGA)BP neural network flight support service time estimation model was established.To verify the performance of the proposed AMGA-BP algorithm,the flight support service time of a large hub airport was taken as the research object,and the traditional GA-BP and BP algorithms were taken as the comparative experiments to estimate the flight support service time.Experimental results show that the AMGA-BP algorithm is more accurate than BP and GA-BP algorithms.