可变编组动车组短期不同席别客流预测研究

针对高速铁路短期不同席别客流做出准确预测,可以提高不同席别客流与动车组编组之间的匹配程度,尽可能做到按需按流配车,实现经济和社会效益最大化。运用熵值法、变异系数法及BP神经网络的组合方法,构建可变编组动车组短期不同席别客流预测模型。该模型以BP神经网络算法为主,通过对往年高速铁路不同线路客流数据进行分析处理,辅以熵值法和变异系数法去除影响高速铁路不同线路席别客流数据的基础因素和随机因素,采用获得的影响系数对BP神经网络运算数据进行修正,以实现短期不同席别客流需求的预测。案例分析表明,该模型在短期不同席别客流预测上具有良好的精确度。

A Research on Short-Term Forecasting of Variable-Combination EMU at Different Seating Classes

Accurate forecasting of short-term passenger flow at different seats on high-speed railway can improve the matching degree between passenger flow at different seats and the formation of EMUs, and maximize economic and social benefits by distributing cars on demand as far as possible. In this paper, a short-term different passenger flow forecasting model for variableform EMUs is constructed by using combination method of entropy method, coefficient of variation method and BP neural network, mainly based on BP neural network algorithm. By analyzing and processing the passenger flow data of different lines of high-speed railway in previous years, the entropy method and coefficient of variation method are used to remove the basic factors and random factors that affect the passenger flow data of different lines of high-speed railway. Through the obtained influence coefficient to modify the BP neural network operation data to predict the short-term different passenger flow demand. The case analysis shows that the model has good accuracy in short-term passenger flow forecasting at different seats.