黄裕春, 贾巍, 雷才嘉, 方兵华, 刘涌, 李洋洋. 基于混沌多目标蚁狮优化算法和核极限学习机的冲击性负荷预测模型[J]. 现代电力, 2023, 40(6): 1043-1051. DOI: 10.19725/j.cnki.1007-2322.2022.0105
引用本文: 黄裕春, 贾巍, 雷才嘉, 方兵华, 刘涌, 李洋洋. 基于混沌多目标蚁狮优化算法和核极限学习机的冲击性负荷预测模型[J]. 现代电力, 2023, 40(6): 1043-1051. DOI: 10.19725/j.cnki.1007-2322.2022.0105
HUANG Yuchun, JIA Wei, LEI Caijia, FANG Binghua, LIU Yong, LI Yangyang. Impact Load Forecasting Model Based on Chaotic Multi-objective Antlion Optimization Algorithm and Kernel Extreme Learning Machine[J]. Modern Electric Power, 2023, 40(6): 1043-1051. DOI: 10.19725/j.cnki.1007-2322.2022.0105
Citation: HUANG Yuchun, JIA Wei, LEI Caijia, FANG Binghua, LIU Yong, LI Yangyang. Impact Load Forecasting Model Based on Chaotic Multi-objective Antlion Optimization Algorithm and Kernel Extreme Learning Machine[J]. Modern Electric Power, 2023, 40(6): 1043-1051. DOI: 10.19725/j.cnki.1007-2322.2022.0105

基于混沌多目标蚁狮优化算法和核极限学习机的冲击性负荷预测模型

Impact Load Forecasting Model Based on Chaotic Multi-objective Antlion Optimization Algorithm and Kernel Extreme Learning Machine

  • 摘要: 针对冲击性负荷预测问题,提出了一种基于混沌多目标蚁狮优化算法(chaotic multi-objective antlion optimization algorithm, CMOALO)和核极限学习机(kernel extreme learning machine, KELM)的冲击性负荷预测模型。首先,为了降低预测难度,使用集合经验模式分解(ensemble empirical mode decomposition, EEMD)将原始冲击性负荷分解为一系列更为平稳的子序列。为了同时提升模型的预测精度和稳定性,提出了一种MOALO;其次,为进一步提高算法的解搜索能力,将MOALO与混沌运算融合,提出了CMOALO算法,将其用于优化KELM。最后通过某地区真实采集的冲击性负荷数据对所提出的EEMD-CMOALO-KELM模型进行验证。通过案例分析可知,所提出的冲击性负荷预测模型,无论是在预测精度还是预测稳定性方面,性能最好。

     

    Abstract: In allusion to the forecasting of impact load, an impact load forecasting model based on chaotic multi-objective antlion optimization algorithm (abbr. CMOALO) and kernel extreme learning machine (abbr. KELM) was proposed. Firstly, to decrease the difficulty of forecasting the ensemble empirical mode decomposition (abbr. EEMD) was utilized to decompose the original impact load into a series of smoother subseries. Secondly, to simultaneously improve the forecasting accuracy and stability of the proposed model, a multi-objective ant lion optimization algorithm (abbr. MOALO) was proposed. Thirdly, to further improve the solution search ability of the algorithm, the MOALO was integrated with chaotic operation to put forward CMOALO algorithm and applying the latter to optimize KELM. Finally, the put forward EEMD-CMOALO-KELM model was verified by true-collected impact load data in a certain region. It can be know by case study that the proposed impact load forecasting model possesses the best performance in both aspects of forecasting accuracy and stability of predicted results.

     

/

返回文章
返回