Functional Link Artificial Neural Network with Cloud Estimation of Distribution Algorithm for Traffic Flow Forecast

The traditional artificial neural network traffic flow forecast model based on gradient descent algorithm has the problems of slow convergence and local optimum. In order to improve the forecast accuracy of traffic flow, a forecast model based on functional link artificial neural network(FLANN) and cloud estimation of distribution algorithm(CEDA) is proposed. In the model, CEDA inspired from cloud model is used to optimize the initial connection weights of FLANN. It firstly extracts the global statistical information about the search space and builds a cloud model of promising solutions by the backward cloud generator. Then, new solutions are generated from the cloud model by the forward cloud generator, and the best solutions are selected from current solutions and new solutions to form next population. The algorithm is applied to the empirical traffic flow forecast. The experimental results show that the model is effectiveness and feasibility, and has a higher forecast accuracy compared with the traditional ANN forecast model with gradient descent algorithm.

[1]  Yanbo Li,et al.  Highway Traffic Flow Nonlinear Character Analysis and Prediction , 2015 .

[2]  Tao Cheng,et al.  An Integrated Cloud Model for Measurement Errors and Fuzziness , 2006 .

[3]  Michael J Demetsky,et al.  SHORT-TERM TRAFFIC FLOW PREDICTION: NEURAL NETWORK APPROACH , 1994 .

[4]  Maurice Clerc,et al.  The particle swarm - explosion, stability, and convergence in a multidimensional complex space , 2002, IEEE Trans. Evol. Comput..

[5]  Xiang Li,et al.  A New Performance Evaluation Method for Automatic Target Recognition Based on Forward Cloud , 2007 .

[6]  Billy M. Williams.  Multivariate Vehicular Traffic Flow Prediction: Evaluation of ARIMAX Modeling , 2001 .

[7]  Wei-Hua Lin.  A Gaussian maximum likelihood formulation for short-term forecasting of traffic flow , 2001, ITSC 2001. 2001 IEEE Intelligent Transportation Systems. Proceedings (Cat. No.01TH8585).

[8]  Y. Takefuji,et al.  Functional-link net computing: theory, system architecture, and functionalities , 1992, Computer.

[9]  Yu A-long.  Research on the dynamic compensation based on genetic wavelet neural networks for the robot wrist force sensor , 2010, The 2nd International Conference on Information Science and Engineering.

[10]  Xuemei Shi,et al.  Uncertainty reasoning based on cloud models in controllers , 1998 .

[11]  Ehab Al-Shaer,et al.  Using Online Traffic Statistical Matching for Optimizing Packet Filtering Performance , 2007, IEEE INFOCOM 2007 - 26th IEEE International Conference on Computer Communications.

[12]  Joe Whittaker,et al.  TRACKING AND PREDICTING A NETWORK TRAFFIC PROCESS , 1997 .

[13]  Huang Zhong-xiang.  Short-Term Traffic Flow Prediction Based on RBF Neural Network , 2004 .

[14]  Lalit M. Patnaik,et al.  Adaptive probabilities of crossover and mutation in genetic algorithms , 1994, IEEE Trans. Syst. Man Cybern..

[15]  Haijun Wang,et al.  Spatial Clustering Method Based on Cloud Model , 2007, Fourth International Conference on Fuzzy Systems and Knowledge Discovery (FSKD 2007).