中文摘要：

提出了基于多442（Multi．AgentsSystem，MAS）＋#N的电力负荷预测模型，将不同的预测模型封装在不同的Agent中，利用各Agent之间的通信、合作机制来共同完成预测问题的求解，以近似日模型、回归模型、神经网络模型和规则修订模型为基础，进行协同预测。并通过实验证明了协同预测模型不仅能达到较好的预测精度，同时预测结果相对稳定，对提高实际负荷预测精度具有很现实的意义。

英文摘要：

This paper proposes the electricity load forecasting model based on the multi-agent（Multi-Agents Sys- tem, MAS） coordination, and encapsulates different forecasting models into different Agents, by using communica- tion between the Agents and the corporation mechanism, the forecasting problem can be solved, and it cooperatively forecasts based on the approximate day model, regression model, neural network model and rules revision model. The experimental results show that the cooperative forecasting model can get better forecasting precision, and the forecasting results are relatively stable, which is significant for the improvement of real load forecasting precision.