Heat Load Forecasting Based on Improved AGA-BP Non-linear Combined Model

A new combined BP neural network model based on accelerating genetic algorithm is put forward in this paper. On the foundation of traditional BP neural network, this method is given better iteration values improved by accelerating genetic algorithm, thus and increase iteration rate and avoid sinking into local minimum. Then, it is applied to forecast the heat load in a certain area, and compared with other forecasting methods. The calculation sample shows the exactitude and efficiency of this combined forecasting model.