Research Of Electric Power Short-Term Load Forecasting Based On Information Entropy And Genetic Neural Network

For a multifactor power load prediction problem, this paper attempts to propose anew method for Short-Term Load Forecasting (STLF), by combining rough set and artificialneural network. The conception of information entropy in rough set is employed to reducefactors of loads and input variables of the input layer;genetic algorithm is used to optimizethe initial network parameters. Meanwhile, several weather factors (Effective Temperature,Temperature Humidity Index, Comfort Index ), which reflect the effort oftemperature ,humidity and wind power on human, are introduced to evaluate the change ofSTLF under weather condition. The implemented program based on the proposed method isused for the STLF in the actual network, the testing result illustrate that the forecastingaccuracy is satisfactory, accordingly it shows the validity and practicability of the method.