Rail transit short-term passenger flow prediction method and system based on graph convolutional neural network

Abstract

The invention discloses a rail transit short-term passenger flow prediction method and system based on a graph convolutional neural network. The method comprises the following steps: 1, ; representingthe urban rail transit network by using a graph structure G = (V, E, A); representing historical passenger flow conditions of m time intervals before the t moment by using a characteristic matrix F =(Xt, Xt-1, Xt-2,... Xt-m + 1) at the t moment; wherein V represents the number of subway stations, E edges exist between the stations, A belongs to Rn * n and is an adjacent matrix representing whether the stations are adjacent or not, the dimension of F is n * m, and n is the number of the stations; and constructing a deep learning model, and learning a mapping relationship between the feature matrix F and the adjacent matrix A at the moment t and the predicted passenger flow condition of each station at the moment t + 1. The method can accurately predict the short-time passenger flow condition of the rail transit, and has important guiding significance for the whole-network-level short-time passenger flow monitoring, real-time management and engineering practice of urban rail transit operation management departments.

Description

technical field [0001] The present invention relates to the technical field of rail transit passenger flow analysis, and more particularly, to a short-term passenger flow prediction method and system for rail transit based on a graph convolutional neural network. Background technique [0002] With the rapid development of urban rail transit network, passengers have put forward higher requirements for the operation management and service level of urban rail transit. In order to effectively improve the modernization and intelligence level of rail transit operation management, it is very important to accurately grasp the time and space distribution of passenger flow in the whole network before and after occurrence. [0003] Throughout the development of short-term passenger flow forecasting, it can be roughly divided into three stages. The first stage is a traditional model based on mathematical statistics, the second stage is a model based on machine learning, and the third s...

AI Technical Summary

Problems solved by technology

[0007] In short, all current short-term passenger flow forecasting models mainly have the following problems: traditional models based on mathematical statistics have problems such as poor real-time performance and low prediction accuracy; although machine learning-based models have improved the short-term However, the influence of the spatio-temporal characteristics of the entire network passenger flow on the prediction results is not considered in the prediction process, and most of these models only predict one or a few stations, and it is impossible to use one model for the entire network. All stati

Images