The Research And Application Of Swarm Intelligence Algorithm To Optimize RBF Neural Network

As a branch of the meta heuristic algorithm of swarm intelligence algorithm,in recent years by scholars attention.A lot of research had been made in terms of theory and application,especially the swarm intelligence optimization algorithm and artificial neural network combined with the formation of hybrid neural network model,has been widely used in the field of practical engineering,such as signal and image processing,pattern recognition,function optimization,prediction model.Under the background of global warming and frequent haze in our country,the development of new energy has become the current trend.At present,the development of new energy is mainly wind and photovoltaic.The randomness and intermittent characteristics of wind speed limits the proportion of grid connected wind power,in the areas of China where new energy is rich had occurred serious abandoned wind and abandoned light phenomenon.Therefore,the accurate forecasting of the wind speed of wind resource assessment,wind farm making the plan of power generation and improve the proportion of wind power grid,has the important practical significance.In this paper,RBF neural network and swarm intelligence algorithm optimize RBF neural network research status at home and abroad are discussed,and the details of the RBF neural network,PSO algorithm and CSO algorithm.We constructed PSO-RBF,CSO-RBF two hybrid model,and based on the improved CSO algorithm to build a new hybrid model of ? CSO-RBF.Under MATLAB platform programming the three hybrid model,and on average the measured wind speed data of a wind farm in Hebei province were carried out empirical analysis.The experimental results show that the three hybrid model prediction accuracy and stability are better than the single RBF neural network model,the CSO-RBF model in the stability is close to PSO-RBF,but the accuracy of the forecasting model of CSO-RBF is better than PSO-RBF,and the prediction accuracy and stability of ? CSO-RBF is superior to PSO-RBF model and CSO-RBF model.