Research On Ls-svm Short-term Load Forecasting Based On Artificial Bee Colony Optimization

Short-term load forecasting is a very important work in power department, it is related to power generation units to the reasonable arrangement of power generation plan, so it is necessary to study. Accuracy of result of the short-term load forecasting affects the plan of the power sector and safety of power system and steady operation. Therefore, using appropriate short-term load forecasting method and the maximum increasing accuracy of forecasting has become an important research topic in the field of load forecasting.At first,This paper introduces the research background and significance of load forecasting and makes classification of load forecasting, explains the characteristics of short-term power load forecasting, methods and research status, etc. Secondly, combining with the characteristics of power load, the paper puts forward forecasting method based on least squares support vector machine (SVM). The model parameter has a great influence on the results of foreasting, so the optimization of least squares vector machine parameters is very important, this paper take use of artificial colony algorithm and the improved colony algorithm to optimize, illustrate principle and process of two kinds of algorithms, and verifies the performance of two algorithms. Finally, building the LS-SVM, the ABC-LSSVM and CABC-LSSVM forecasting model separately, according to load forecasting data of liaoning, and comparing the forecasting results of each model, combining with the influencing factors of load forecasting,analyses and concluds that forecasting results verified that the artificial colony algorithm optimization is effective, which based on improved swarm optimization LSSVM forecasting precision is highest.