Short-Term Load Forecasting Based on Improved Gene Expression Programming

Short-term load forecasting is an important part of the modernized power system administration. Some methods have been applied to short-term load forecasting and obtained a certain achievement. Considering the characteristics of gene expression programming (GEP), it is possible to apply GEP to short-term load forecasting. But there still are some shortcomings of GEP. Such as the initial population that is generated randomly, mutation rate that can´t be adjusted by itself and evolution result got before that can ´t be utilized. In order to overcome these shortcomings, GEP was improved (IGEP) in the aspects of excessive multiplication, self-adaptive mutation rate and adopting mathematical model got before is proposed. And the IGEP was applied to short-term load forecasting. Firstly, the load series of the same time but different days are chosen as the training samples. Secondly, the load samples are filtered and processed generally. And finally, the short-term load is forecasted classified by weekday and weekend after eliminating the pseudo-data. Compared with the results forecasted by means of GP and GEP, it proves that the method of IGEP in short-term load forecasting is better.