应用信息熵原理的多路径负荷预测协同方法

Method of Multi-path Load Forecasting Collaboration Using Maximum Entropy Principle

将最大信息熵原理应用于多路径负荷预测中,提出了一种对含不确定信息的多路径负荷预测结果的协同方法。该方法将负荷预测结果的期望、二阶中心矩作为约束信息,利用最大信息熵原理得到预测结果的概率分布函数,并应用概率论自动得到综合两条预测路径的高、中、低方案;其次通过两两组合融合直至每组原始数据至少使用一次,不同组合方式结果的信息熵不同,选取信息熵最大的一组为最终预测方案。算例表明,应用上述方法所得预测结果的信息熵总是大于原始预测结果的信息熵,熵的提高验证了该方法在负荷预测信息融合中的有效性。

Applying maximum entropy principle in the process of multi-path load forecasting, a collaborative method is proposed to solve the problem that information conflicts with uncertainties in load forecasting. Firstly, taking the ex- pectation and second order central moment of forecasted results into account and using the maximum entropy principle, this model receives the probability distribution function of the forecasted values and obtains the final high, middle and low forecasting schemes of any two forecasting paths. Secondly, this forecasting method makes two groups of data inte- gration until each group is used more than once. Because different entropy of forecasting results are decided by different combination mode, the final forecasting scheme can be chosen according to the value of entropy. Case studies have shown that entropy of the collaborative forecasting result is greater than that of the original one, demonstrating the ef- fectiveness of information fusion of this method.