Improvement Of Gravity Model For Prediction Of Traffic Distribution

The traffic distribution forecast in urban traffic planning is an important step in traffic demand forecasting.At present,there are many kinds of traffic distribution prediction models,of which the most widely used one is the gravity model.However,the traditional gravity model only considers the factors of the residents' travel distance,and the prediction results tend to have large deviations from the actual ones.Therefore,in order to solve the problems existing in the practical application of the traditional gravity model,an improved model with higher accuracy is proposed.This article firstly re-analyzed the survey results of Shijiazhuang residents' travel,and analyzed the problems existing in the practical application of the traditional gravity model in conjunction with the investigation situation,and analyzed the existing problems in detail.Based on this,the idea of improvement and the framework of distributed forecasting method based on the classification of land use properties are proposed.Then the corresponding improved model is constructed and the forecasting problem of traffic generation in the improved model is studied.Finally,the improved model was used to reanalyze the survey results of Shijiazhuang residents' trips,and the effectiveness of the improved models was verified.The results show that this prediction method effectively improves the accuracy of the traditional gravity model.