中文摘要：

利用粗糙集理论处理大数据量、消除冗余信息等方面的优势，找到了与负荷直接相关的因素。以利用遗传程序设计进行演化建模对贵州电网日96点负荷实例进行了预测，与BP神经网络法相比，本模型预测精度高，在短期负荷预测中具有有效性和可行性。

英文摘要：

Using the advantages of RS theory for processing large data and eliminating redundant information, it finds the relevant factors to load. The forecasting model is established by means of GP evolutional algorithm. It is applied to short-term load forecasting using the actual daily load 96-points data for GuiZhou power grid. The results demonstrate that the precision of the proposed forecasting models is better than that of BP and the proposed model is feasible and effective for short-term load forecasting.