Bacterial foraging optimization based neural network for short-term load forecasting

Because of the nonlinear characteristics of the power system loads, the short-term load forecasting is difficult to realize. The neural network (NN) was employed in this study for forecasting. However, NN is easy to be trapped in local minima and converge too slow. The traditional training methods based on gradient searching technique are not effective and fast. Therefore, bacterial foraging optimization (BFO) was adopted to train the NN. BFO is a novel and powerful global search technique, and it can find the weights/biases of the neural network quickly and accurately. Experiments indicate that the proposed BFO-NN is superior to GA-NN with respect to convergence speed and forecast accuracy.