Research And Application Of Short-term Traffic Flow Forecast Based On Improved BP Neural Network

With the rapid development of the economy and the continuous improvement of the quality of life,the number of motor vehicles in the city is also increasing rapidly,resulting in more and more complicated traffic conditions,and consequently,congestion of traffic roads,continuous occurrence of traffic accidents,and Environmental pollution and other problems caused by automobile exhaust,so more accurate prediction of short-term traffic flow can relieve urban traffic pressure and also better promote the development of intelligent transportation.In this paper,the short-term traffic flow prediction algorithm based on BP neural network is selected as the research object.For the BP neural network in the actual prediction,it is easy to fall into the local minimum,slow convergence,and low prediction accuracy.Improve the BP neural network prediction algorithm,and use the improved algorithm for road short-term traffic flow prediction.The research work mainly includes the following aspects:(1)The establishment of short-term traffic flow prediction and prediction model.First,the relevant information of short-term traffic flow prediction is summarized.When establishing a prediction model,select actual traffic intersections based on the BP neural network.Determine the number of nodes in the input layer of the prediction model by calculating the correlation of different flow sequences.Then select the number of hidden layers and The number of hidden layer nodes finally determines the parameters of the model.(2)Improvement of prediction algorithm based on BP neural network.Aiming at the shortcomings of BP neural network training such as low efficiency,prone tooverfitting,sensitive to initial weight threshold,and easy to fall into local minimum,the BP learning algorithm was firstly adopted by the adaptive momentum estimation(Ada M)algorithm.Improve.In terms of weight threshold optimization,the ant colony optimization algorithm is used to give the optimal weight threshold.Based on the analysis of the existing ant colony algorithm and the BP neural network fusion algorithm,the ant colony optimization algorithm is further improved.Adapting the volatility coefficient,the pheromone iterative elite selection strategy,and the population iteratively adding mutation factors to improve the three aspects,and finally get the optimal solution to complete the task of short-term traffic flow prediction.(3)Application of improved prediction algorithm in traffic flow prediction system.The traffic flow forecasting system is designed,and the improved forecasting algorithm is applied to it.Finally,the system's functional test analysis is performed to confirm the system reliability,which shows that the algorithm in this paper has certain practical value.