Short-term Load Forecasting of Smart Grid Systems by Combination of General Regression Neural Network and Least Squares-Support Vector Machine Algorithm Optimized by Harmony Search Algorithm Method

This paper presents an optimization algorithm to solve the short-term load forecasting problem more quickly and accurately in progress of smart grid development. The new approach employs generalized regression neural network (GRNN) to select influence factors of short-term load, and then a least squares-support vector machine (LS-SVM) based on harmony search algorithm (HS) optimization algorithm was proposed that improving the computing accuracy and speed through a novel category of bionic algorithm, and determining the hyper-parameters of LS-SVM through HS optimization algorithm fleetly and reasonably. Simulations have been made comparing the proposed algorithm with several other algorithms commonly used to solve short-term load forecasting problems. The actual implementation result proves that the proposed algorithm can achieve higher prediction accuracy and better computational speed which is more practical for short term load forecasting.