Simulated Annealing Chaotic Partice Swarm Optimization Algorithm And Its Application In Short-Term Traffic Flow Forecasting

With the prediction of Traffic flow has turned into a hot area of Intelligent Transportation System (ITS), the major cities of China has started to undertake the corresponding strategic planning research on ITS. Due to the nonlinear and uncertainty of urban traffic flow, the traffic predicition which is comprehensive, accurate and timely is of great importance in the Urban traffic control system.In this paper, we have analyzed the predictability of the traffic flow data collected from PEMS. To analyze the chaos of traffic flow further by calculating the maximum Lyapunov index of traffic flow time series. the C-C method is used to calculate the reconstructed phase space embedding dimension m and time delay t.To prove the predictability of the traffic flow deeply by simulating and validating with the measured traffic flow data.Based on the research of neural networks, we have set up BP neural network model which has applied to short-term traffic flow forecasting. The simulating resault has shown that this algorithm has a defect of easy running into partial minimum. it is necessary to improve the BP neural network model. we have introduced the basic particle swarms optimization and some ahybrid algorithms. Based on these algorithms have set up some neural network models, Such as PSO-BP、CPSO-BP、SAPSO-BP to predict the traffic flow. Carrying on a predictive simulation about the traffic data of working days, weekends and holidays with using these above models. The simulation results have shown that, The BP neural network model which has been optimized by Intelligent algorithm is improved in accuracy and convergence greatly. So it can be easier to make a forecast on short-term traffic flow. we can evaluate the advantages and disadvantages of prediction model with the prediction index.In order to make the prediction more prominent, the chaos algorithm has introduced into the simulated annealing particle swarm optimization algorithm, So the Simulated annealing chaotic particle swarm optimization algorithm(SACPSO) has designed. Because the traffic flow data collected at a different time has different influencing factors. So we have set up some prediction models to act simulation experiments which aimed at the traffic flow data from the several ways. After comparative analysis,Research results have shown that SACPSO-BP network traffic flow prediction mode can forecast the traffic flow better.