Application Of BBO Optimization Algorithm In Time Series Prediction

Biogeography-based optimization(BBO)algorithm is a new population-based evolution algorithm.BBO has attracted wide attention from domestic and foreign scholars because of its good global optimization ability and robustness,and has been applied in many real-world optimization problems yet.Time series prediction is closely related to many practical applications in people's life,and it has always been the hot and difficult point in experts' study.There is great important theoretical and practical application value for how to improve the prediction accuracy of time series prediction in engineering applications.The prediction model based on extreme learning machine(ELM)has been widely used in engineering applications,and also achieved good predictive performance.ELM method based optimization algorithm is supposed to be a favorable candidate to improve the accuracy of time series prediction.Focusing on time series prediction,the BBO algorithm is used to optimize the network structure and parameters of ELM,and the BBO-ELM adaptive prediction method which based on BBO algorithm optimized ELM method in this thesis.The main research contents of this thesis are as follows:(1)The basic theory and mathematical model of BBO algorithm,the optimization problems in engineering applications transformed into mathematical model based on BBO algorithm studied,and the problem that how to optimization and implementation of the model is further studied,and the BBO algorithm different from other evolutionary algorithms is presented.The basic concepts and modeling methods of time series prediction are introduced briefly.And the ELM method is tested on the standard chaotic time series,experimental results show that the ELM method has a good ability to predict the nonlinear time series.(2)Aiming at the key points of how to find the valid and necessary historical information in time series,the prediction model based on BBO algorithm and ELM method is studied,and the BBO is used to optimize the set of input variables,the configuration and the parameters of hidden-layer nodes as well as regularization factor.Therefore,the BBO-ELM method is proposed.O n the basis of the proposed method,the cosine migration model and chaotic mapping theory are introduced to improve the performance of MCBBO-ELM method and CBBO-ELM method.The proposed methods are then applied to the prediction of the Machey-Glass chaotic time series,and compared with the other methods such as GA-ELM,etc.under the same conditions.Experimental results confirm that the predic tion performance of BBO-ELM is improved significantly,and its effectiveness also be verified.(3)The proposed methods are applied to the example of network traffic prediction,wind power prediction and traffic flow prediction examples,experimental results confirm that the proposed methods provide superior performance on the convergence speed and prediction accuracy than the contrast methods under the same conditions,confirming the validity and robustness of the proposed methods in this thesis.