基于改进秩次集对变权组合模型的月度负荷预测

Monthly Load Forecasting Based on Variable Weigh Combination Model of Improved Rank Set Pair Analysis

为进一步提高月用电负荷预测精度,本文提出一种基于改进秩次集对和灰色模型的变权组合预测方法。采用改进秩次集对算法,在秩次集对模型中引入天气指标,并利用熵权法确定各指标权重,增强了秩次集对算法的适应性和有效性。接着采用变权法将改进秩次集对模型和灰色模型进行变权组合,不断滚动优化组合模型权重,改善了单一模型预测精度的稳定性。实例预测结果验证了该方法的有效性。

In this paper,a variable weight combination mode based on improved rank set pair analysis(RSPA)and gray model is proposed to improve the accuracy of monthly load forecasting.Firstly,the improved RSPA is proposed,in which the weather index is introduced,and the weights of index are set by the entropy weight algorithm.The method improves adaptability and effectiveness of the RSPA algorithm.Then,the variable weight method is used to combine improved RSPA and gray model with variable weights,and the weight of combination model is optimized continuously.The method improves the stability of the single model forecasting accuracy.Simulation results verify the validity of the proposed method.