Short-term Traffic Flow Prediction Based On Particle Swarm Optimization And Recurrent Neural Network

With the improvement of people’s life,cars have entered millions of households as an indispensable tool,the problem of urban traffic congestion is getting worse.In the face of severe traffic congestion,navigation systems have gradually become the first choice for people to travel.As the theoretical basis of the navigation system,the accuracy of short-term traffic flow prediction has a great influence on it.Accurate traffic flow prediction can not only allow the navigation system to plan a reasonable route,but also provide traffic managers with accurate and real-time road information,so that people can make reasonable responses to possible congestion in advance,so as to avoid large-scale traffic congestion.Therefore,how to realize the accurate prediction of traffic flow has become a hot research topic.This paper mainly studies the short-term traffic flow prediction based on gate neural network(GRU).First of all,the GRU neural network is used to model the traffic flow prediction.At the same time,the basic principles of the particle swarm optimization algorithm are analyzed and researched,Aiming at the shortcomings of particle swarm optimization algorithm,such as easy to appear local extremum and low accuracy,the improved particle swarm optimization algorithm(IPSO algorithm)is proposed then IPSO and GRU are combined for research;next,in order to further improve the accuracy of traffic flow prediction,this paper uses IPSO-GRU and RBF network to predict the same data.And conduct fusion research on their prediction results;finally,we use the IPSOGRU prediction model as the theoretical basis to develop a set of traffic flow prediction display platform.The main work of this thesis is as follows:1.Research on Short-term Traffic Flow Forecast Based on IPSO-GRUThis research uses an improved particle swarm optimization algorithm to optimize the hyperparameter combination of the GRU network based on the GRU-based shortterm traffic flow prediction model.and establishes an IPSO-GRU short-term traffic flow prediction model.And use traffic flow data to verify it to prove its effectiveness.2.Integrate IPSO-GRU and RBF neural network to further improve the accuracy of short-term traffic flow predictionBased on the IPSO-GRU prediction model,combined with the RBF neural network,the two prediction models are weighted and fused to capture traffic flow characteristics from multiple dimensions and make predictions.This method breaks through the limitations of a single model for predicting a single dimension.So as to further improve the accuracy of prediction.3.Short-term traffic flow prediction platform based on Web terminalIn order to better apply the IPSO-GRU prediction model to daily life,this article takes Haixia Road in Nan’an District,Chongqing City as the main research section.A set of short-term traffic flow prediction and display platform has been developed using the SSM framework.We use manually collected data as the research basis to further verify the effectiveness of our display platform.