基于算法融合的自适应短期负荷组合预测模型研究

Research on self-adaptive integrated model for short-term load forecasting based on algorithm combination

组合预测把多种单一预测方法按一定方式结合,综合利用各种预测方法所提供的信息,并在综合这些信息的基础之上进行最优组合。采用支持向量机(SVM)实现分时段变权重组合预测,描述多种方法的预测结果与实际负荷的非线性关系,并采用改进粒子群(PSO)与模拟退火(SA)自学习融合的协同优化方法SA-MPSO对SVM模型参数进行优化,用两种不同特性的测试函数对该优化算法的收敛性进行测试,通过多次测试平均值验证其收敛性。实例仿真中,SA-MPSO优化的SVM模型实现对三个不同预测模型的组合,预测结果表明,该方法除了避开传统组合预测模型权重复杂求取问题,且参数优化自适应能力强,有利于预测精度的提高。

Combination forecasting model integrates single prediction method according to a certain mode to utilize the information provided by various methods. This paper adopts support vector machine to implement the time-phased varying weight integrated forecasting and to describe nonlinear relationship between the results of different methods and the actual data. A new algorithm SA-MPSO which is based on improved particle swarm optimization and simulated annealing is also put forward to optimize the parameters of support vector machine. The convergence test of this optimization algorithm is studied by using two functions with different characteristics. The results show that the proposed method can avoid the complex weight calculation of the traditional model and has strong self-adaptive ability of parameter optimization, which is helpful to improving the forecasting accuracy.