Adaptive neural networks for tariff forecasting and energy management
Herman Wezenberg, M.B. Dewe · 2002
The paper looks at using a hybrid combination of recurrent neural networks trained with a temporal difference procedure for predicting local power tariff rates and energy use, with the intent of cost-effectively utilising electric power to heat the water in, for example, domestic hot water cylinder. The neural networks are adaptive and capable of both linear and non-linear time series forecasting with a minimum of training data.