Load Forecasting Based On Analytic Hierarchy Process And Radial Basis Function Neural Network

The using of the land doesn't change basically and the land is developed stable and mature in a relatively well-developed old city, so the paper established a classification and long-term load forecasting model based on analytic hierarchy process (AHP) and radial basis function neural network (RBFNN) for this area, which provides a data base for getting the spatial area load. Using the decision-making analysis ability for multi-criteria and multi-objective problems of AHP, we considered many factors, combined the advantages of combination forecasting method, and established the AHP evaluation system of three indicators about the total load, the increase in it and the growth rate respectively, to get the optimized model. Also, we combined this with the advantages of the RBFNN, that is, the approximation capability, the simple structure and the fast learning, then taking the predictive value of the optimal model and the related factors as the input of it to get the fitted load forecast. At last, on basis of AHP and RBFNN, the paper made classification and long-term load forecasting in a practical area, we considered many factors, improved the accuracy of the prediction, and demonstrated the superiority of the proposed model in distribution load forecasting.