With the development of the economy, intelligent transportation plays an important role in traffic control and induction. Traffic flow prediction is the basis of them.But the traditional technology of traffic flow prediction can not meet the actual requirements, many new prediction methods, such as intelligent computing, has been researched and applied in the field.The neural network based on particle swarm optimization was studied and applied in the traffic flow prediction in this paper. The contents were organized as follows:(1) Based on the existing short-term traffic flow prediction model, two short-term traffic flow prediction models of information fusion were proposed. The first was a fusion of multiple models with different weighted average, the other was a model combining the complementary advantages of various algorithms.(2) The wavelet neural network was rap laced the traditiona BP neural network, to overcome its withdraws, such as long training time, slow convergence speed, easily trapped into focal minimum value, etc. The simulation results proved that the forecast result was clearly improved.(3) A improved particle swarm optimization algorithm was applied to optimize the weights and threshold values of the wavelet neural network. Then a predictive simulation based on an adaptive mutation particle swarm optimization was founded. The simulated results verified that the convergence and forecast accuracy the wavelet neural network model were improved.(4) The short-term traffic flow prediction algorithm based on chaos particle swarm optimization wavelet neural network were proposed.The new model could help the inert particles escaped from a local extreme point.The simulation results showed that the forecast the fusion model forecast was better than particle swarm optimization wavelet neural network. and the convergence rate was accelerated. |