基于免疫加权支持向量机方法的短期负荷预测

An immune weighted support vector machines approach for short-term load forecasting

在对支持向量机(SVM)方法进行分析的基础上,提出了一种免疫加权支持向量机(IWSVM)方法来预测电力系统短期负荷。其中根据各样本重要性的不同,引入了加权支持向量机方法,然后利用免疫规划算法对其进行参数优化。免疫规划算法利用浓度和个体多样性保持机制进行免疫调节,有效地克服了未成熟收敛现象,提高了群体的多样性。电力系统短期负荷预测的实际算例表明,与支持向量机方法相比,所提免疫加权支持向量机方法具有更高的预测精度。

The paper presented an immune weighted support vector machines （IWSVM） method for short-term load forecasting. The weighted support vector machines method was introduced to express the differences among the training samples. The immune programming algorithm, inspired by the immune system of human and other mammals, was used to optimize the parameters of weighted support vector machines. The algorithm has the advantage in preventing premature convergence and promoting population diversity. The forecasting results demonstrate that the proposed method has higher forecast precision for short-term load forecasting.