Self organization and learning methods in short term electric load forecasting: A review

Abstract

This paper reviews unconventional methods used in short-term load forecasting. The basic theory of these methods and their suitability to short term load forecasting is discussed. Application and the basic formulation strategy adopted for the purpose are also discussed. These methods are classified into supervised and unsupervised learning and self-organizing with optimization categories. Different models of artificial neural network, fuzzy logic, evolutionary programming, simulated annealing, learning machine, and expert system have been dealt within appropriate classification of each.