The Improvement Of Gene Expression Programming And Its Application In Load Forecasting

Gene Expression Programming(GEP)algorithm is a new technology of searching and knowledge discovery on the basis of the inheritance and development the advantages of genetic algorithm and genetic programming,borrows the gene expression rules of biological inheritance.The algorithm has attracted much attention from domestic and foreign researchers because of its strong point of stronger ability to solve problems,easy to carry on genetic operation and unique coding method.It has been widely used in typical applications of research areas of function mining,prediction,association rules and classification clustering,includes time series and data flow prediction model,disaster analysis,multi-layer association rule mining,feature extraction,mining recursive function,etc.The main research work of this thesis is to improve the standard GEP algorithm,apply the improved algorithm to the field of power load forecasting,and construct the power load forecasting model based on GEP algorithm.On this basis,the power load forecasting prototype system based on gene expression programming is designed.The specific work of this thesis mainly has the following points:(1)Based on the idea of niche,this thesis improves the standard GEP algorithm and gets a gene expression programming by using a niche technology of outbreeding fusion(OFN-GEP).(2)In this thesis,OFN-GEP algorithm has applied to the forecasting of electric load,taken the electricity load of residents in Xianyang City as the research object,constructed the power load forecasting model based on OFN-GEP algorithm.The OFN-GEP algorithm is compared with similar to the improved algorithm of the latest literature,the superiority and the high prediction accuracy of the power load forecasting model based on the OFN-GEP algorithm is verified.(3)In this thesis,the power load forecasting prototype system is designed and provides a scientific and effective reference for the power plant dispatcher.The power load forecasting system based on gene expression programming algorithm in this thesis has played a practical role in a power plant in Xianyang City,which provides some reference value for the dispatcher of power plant.