Dynamic forecasting of electric load consumption using adaptive multilayer perceptron(AMLP)

Abstract

Electric energy plays a vital role in the achievement of social, economic and environment development of any nation. Thus, efficient demand planning and production of energy is needed to avoid too much over/under-estimation of electric load. In this study, the researchers proposed a scheme with eight steps for a dynamic time series forecasting using adaptive multilayer perceptron with minimal complexity. Two different data sets; each divided into three overlapping parts (training, validating and testing sets), from two different countries were used in the experiments to measure the robustness and accuracy of the models produced by the AMLP. Experiments results show the effectiveness of the proposed scheme for AMLP in forecasting the electric load consumption based on the calculated coefficient of variance of RMSD, CV(RMSD).