Air traffic flow of genetic algorithm to optimize wavelet neural network prediction

The scientific and accurate forecast of air traffic flow is not only an effective protection to maintain the air traffic flow continued and unimpeded, and also is an important basis for the air traffic flow management(ATFM) to make decisions and development strategies. Based on the character of flow prediction, the prediction method of genetic algorithm to optimize wavelet neural network is proposed. It uses genetic algorithms with the natural evolution laws to conduct the pre-optimized training for the connection weights and stretching translation scales of the wavelet neural network, overcoming the drawbacks of easy to fall into local minima and causing oscillation effect of wavelet neural network with a single gradient descent method. The air flow prediction simulation using the GA-WNN prediction model demonstrates the validity of the model.

[1]  F.H.F. Leung,et al.  Genetic algorithm-based variable translation wavelet neural network and its application , 2005, Proceedings. 2005 IEEE International Joint Conference on Neural Networks, 2005..

[2]  Lei Jia,et al.  Wavelet network with genetic algorithm and its applications for traffic flow forecasting , 2004, Fifth World Congress on Intelligent Control and Automation (IEEE Cat. No.04EX788).

[3]  Zhao Yong-mei.  The method and application of time series prediction based wavelet neural network , 2004 .

[4]  Ding Yong.  Extended wavelet neural network structure and its optimal method , 2003 .

[5]  Shu-Ching Chen,et al.  Function approximation using robust wavelet neural networks , 2002, 14th IEEE International Conference on Tools with Artificial Intelligence, 2002. (ICTAI 2002). Proceedings..

[6]  Tan Guo.  Research of generalized neural network and it′s application to traffic flow prediction , 2002 .