Short-term load forecasting based on complexity science theory

As a typical and special complexity gigantic system, the power system is facing the challenge from complexity science in the aspects of load forecasting and its management. Therefore, on the basis of complex system theory, a new method used for predicting the short-term load is proposed by means of a series of subsystems divided according to the different areas and types of regional electricity. Support vector machine forecasting model is applied to each subsystem and the results show this model is better than one of neural network in forecasting accuracy.