基于Shapley值的组合预测方法

Combined forecasting method based on Shapley value

负荷预测对电网的安全运行起着至关重要的作用,预测误差较大则从长期和短期来讲都使电网公司面临很大的风险,所以必须提高负荷预测的精度、减小误差,组合预测能够有效地提高预测精度,但是其权重的设定是一个比较困难的问题。应用shapley值方法进行研究,根据分摊的误差来确定各个成员的权重,算例验证表明Shapley值方法确定的组合预测模型可以比其他方法更好地减小预测的误差。

The load forecasting plays a most important role in the safe operation of the power grid. A bigger error in forecasting will bring huge risk to the power grid company in both long and short terms, so the precision of the load forecasting should be improved and the error reduced. The combined forecasting can improve the forecasting precision effectively, however, its weight is relatively difficult to set. Applying the Shapley value method, the weight of every member can be determined according to the distributed error. The calculation example shows that the combined forecasting model determined by the Shapley value method can dramatically reduce the forecasting error.