Short-Term Load Forecasting Based on LS-SVM Optimized by Bacterial Colony Chemotaxis Algorithm

摘要

Aiming at improving the accuracy and speed of short-term load forecasting (STLF), the proposed BCC-LS-SVM model is presented, among which bacterial colony chemotaxis (BCC) optimization algorithm is used to determine hyper-parameters of least squares support vector machine (LS-SVM). BCC is a novel category of bionic algorithm, which takes advantage of the bacterium's reaction to chemoattractants to find the optimum. The algorithm not only has strong global search capability, but also is easy to implement. Thus, BCC is suitable to determine parameters of LS-SVM. Finally, load forecasting examples are used to illustrate the performance of proposed model. The experimental results indicate that the BCC-LS-SVM method can achieve higher forecasting accuracy and faster speed than artificial neural network and LS-SVM with gird search. Therefore, the BCC-LS-SVM model is suitable for short-term load forecasting.