关键词: 交通运输工程；高铁快运；灰色-BP神经网络；快运量预测；高铁分担率

Abstract: With the advancement of green logistics, the “high-speed rail + express delivery” freight model has demonstrated growing application value, and the accurate prediction of its transportation volume holds practical significance for optimizing railway transportation capacity. Taking the Gansu-Qinghai-Ningxia region as an example, the grey relational analysis method was employed to screen out key influencing factors from 15 initial indicators, and a combined prediction model based on grey-BP neural networks was constructed. The research results show that compared with GM (1,1) and ARIMA models, the combined model shows a significant improvement in prediction accuracy. An inverted U-shaped curve characterizing high-speed rail's sharing rate changing with transportation distances was revealed by establishing a Logit model incorporating seven dimensions such as economy, timeliness and environmental protection. And the advantageous transportation distance range is 400~1 000 km. Based on these models, the predicted values of high-speed rail express transportation volume for this region over the next three years are calculated out, providing data support for the allocation of transportation capacity for the railway department.

Key words: traffic and transportation engineering; high-speed rail express transportation; gray-BP neural network; express transportation volume prediction; high-speed rail sharing rate