Research Of Combined Model In Railway Passenger Flow Forecasting Based On Grey Theory And Intelligent Algorithm

Nowadays, with our country modernization and rapid development, railway network gradually developed under the national economy, productivity layout, industrial structure and reasonable division of labor. Inter-city railway as the city's comprehensive transportation system promotes the development of urban industrial economy and the overall ascension of image. However, the forecast of the passenger flow is the basis for the construction project and project research, it is the symbol of economic rationality and important to determine the national construction project. Therefore, to explore the prediction model of the inter-city railway passenger flow and improve the precision of prediction has become an academic topic needs to be further studied.Based on the above background, according to the characteristics of passenger flow of complicated variety, this paper on the basis of previous studies to select the Lanzhou to Zhongchuan inter-city railway as the research object, putting forward the new combined forecasting method, which is combining of grey theory and intelligent algorithm. First of all, through the grey correlation method analyses the relationship between passenger flow and influence factors, choosing the main factors, investigating distribution along the railway traffic, collecting traffic data each year;Then introduces the GM(1, 1) model of grey theory and the theoretical basis of genetic algorithm to optimize neural network model, and builds the linear time series of grey GM(1, 1) model, the influence factor of nonlinear GA- BP model; Then, in order to avoid a single forecasting model to predict the results respectively, the GM(1, 1) model and GA- BP model become combined together to open up a new method of passenger flow data research, which is the combination of three kinds of series, parallel, embedded combination model. And single models and combined models for passenger flow will be illustrated by Lanzhou to Zhongchuan airport inter-city railway. Followed by the mean square error(MSE), the normalized mean square error(NMSE), mean absolute percentage error(MAPE) and R value evaluation index will be compared with the single models and combined models, drawing combined models prediction accuracy is higher than single forecasting models.Meanwhile, in these three combination models, it is concluded that the embedded combination model shows the best prediction effect, giving full consideration to the linear and nonlinear characteristics of passenger flow data has better prediction performance; Finally, through the investigation to collect Lanzhou to Zhongchuan highway transportation fares and the passenger travel intention questionnaire, calculated from the passenger travel cost, time value and so on, using the Logit model to work out the share rate of railway passenger flow, turning the data into the Lanzhou to Zhongchuan inter-city railway traffic.Therefore, according to the example in this research can be concluded that the embedded combination model compared with the single models to forecast passenger flow has a higher feasibility of application, establishing scientific and reasonable operating strategy for railway operation strategy.