Application of SVR with chaotic GASA algorithm in cyclic electric load forecasting
| Autor: | Yucheng Dong, Jing-Tian Sung, Wei-Chiang Hong, Wen Yu Zhang, Guo-Feng Fan, Gary Tsai |
|---|---|
| Rok vydání: | 2012 |
| Předmět: |
Engineering
Electrical load business.industry Mechanical Engineering Evolutionary algorithm Chaotic Building and Construction Pollution Hybrid algorithm Industrial and Manufacturing Engineering Support vector machine General Energy Local optimum Convergence (routing) Autoregressive integrated moving average Electrical and Electronic Engineering business Algorithm Civil and Structural Engineering |
| Zdroj: | Energy. 45:850-858 |
| ISSN: | 0360-5442 |
| DOI: | 10.1016/j.energy.2012.07.006 |
| Popis: | The electric load forecasting is complicated, and it sometimes reveals cyclic changes due to cyclic economic activities or climate seasonal nature, such as hourly peak in a working day, weekly peak in a business week, and monthly peak in a demand planned year. Hybridization of support vector regression (SVR) with chaotic sequence and evolutionary algorithms has successfully been applied to improve forecasting accuracy, and to effectively avoid trapping in a local optimum. However, it has not been widely explored to employ SVR-based model to deal with cyclic electric load forecasting. This paper will firstly investigate the potentiality of a novel hybrid algorithm, namely chaotic genetic algorithm-simulated annealing algorithm (CGASA), with an SVR model to improve load forecasting accurate performance. In which, the proposed CGASA employs internal randomness of chaotic iterations to overcome premature local optimum. Secondly, the seasonal mechanism will then be applied to well adjust the cyclic load tendency. Finally, a numerical example from an existed reference is employed to compare the forecasting performance of the proposed SSVRCGASA model. The forecasting results show that the SSVRCGASA model yields more accurate forecasting results than ARIMA and TF-e-SVR-SA models. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full Text from ScienceDirect