GA based optimized LS-SVM forecasting of short term electricity price in competitive power markets

Development of competitive power markets has emerged an increasing tendency to forecast the future prices among both the producers and the consumers with the major aim of profit maximization. A least-square support vector machines approach, in combination with a hybrid genetic algorithm based optimization is proposed in this paper to forecast market clearing prices in two different power markets. The forecasting results are compared to a select variety of previously proposed methods such as MLP, ARIMA, wavelet- ARIMA, neuro-fuzzy and time series based models. The performed comprehensive comparison demonstrates the remarkable accuracy and effectiveness of the proposed method.