Different ANN Models for Short-Term Electricity Price Forecasting

Abstract

Application of ANN to the electricity price forecasting is gaining importance in the context of electricity markets. This paper presents four different ANN models used for forecasting the electricity price. They are the classical back propagation neural network model, radial basis function neural network model, genetic algorithm based neural network and a direct non-iterative state estimation based neural network model. A case study is made with the downloaded data of the day-ahead pool market prices of the Iberian Energy Derivatives Exchange using the above four different ANN models and the results are compared.