Short-term Traffic Flow Prediction Based On Improved Particle Swarm Optimization BP And RBF Combination Algorithm

Traffic problem in modern life has become a problem which can not be avoided in people’s life. For most people, the traffic conditions have a direct impact on their quality of life and happiness index. In modern life, the traffic problems not only plagued people living in the city, even in some of the towns and villages also appeared the phenomenon of traffic congestion, traffic system is facing more and more challenges, solving the problem of traffic congestion is imminent. As the main means of improving traffic, intelligent traffic plays an important role in improving the traffic jam. The intelligent and informative road traffic is an important part of realizing intelligent traffic, it can provide real-time and accurate traffic flow forecasting and provides a practical and effective source of information, ensures the smooth road. Not deterministic, stochastic and time-varying is the three major characteristics of traffic flow. Therefore, how to establish real-time and accurate short-term traffic flow forecasting model has become the hot spot and difficulty in the study of intelligent transportation system (ITS).In this paper, a road traffic flow data collected in Guiyang as the base of the traffic flow data for screening, remove the data between nine clock at night until the next morning six clock, to ensure the data is useful data to improve traffic.RBF neural network and BP neural network are widely used in traffic flow prediction. In the traffic prediction, in order to improve the prediction accuracy of traffic flow, this paper uses the combination forecasting model. Two methods are used to improve the global searching ability of particle swarm optimization algorithm, which can improve the searching range of BP neural network and the convergence speed and precision of RBF neural network. Combining these two improved prediction algorithms, the combination forecasting algorithm is formed. Experimental results show that based on Improved Particle Swarm Optimization Algorithm of BP neural network and RBF neural network traffic flow forecasting not only has good prediction ability, and applied to the short-term traffic flow forecasting is effective and feasible.