【作者】 屈利； 苑津莎； 张卫华； 李丽；

【Author】 Qu Li, Yuan Jinsha, Zhang Weihua, Li Li(School of Electronic and Information Engineering, North China Electric Power University,Baoding 071003, China)

【机构】 华北电力大学电气与电子工程学院；

【摘要】 人工智能中的事例推理是解决短期负荷预测问题的新方法。事例推理的特点是将历史经验作为代替规则专家系统的"知识"存储在事例库中。对于新的预测问题,系统从事例库中检索出相似事例,并将数据(来自相似事例集)输入到神经网络中训练,这样既有效克服了大数据量训练的盲目性,大大减少了网络输入的节点数,提高训练效率,又提高了预测精度。在事例修正中,针对非正常日的预测误差较大的问题,文中给出了统一的修正公式。经实例验证比较表明,该方案是有效可行的。更多还原

【Abstract】 Case-based reasoning (CBR) system pertained to artificial intelligence (AI) is a novel method to solve short-term load forecasting (STLF) problem. The characteristic of CBR system is that the histories experiences are utilized for forecast, which replace "the knowledge" in the rule-expert-system, and preserved in the cases base. As to the new problem to be forecasted, the system searches the similar cases from the cases base and input the datum (from the similar cases set) into neural network for training. On the one hand, effectively, it hurdles the blindness of a large scale of datum training, reduces the nodes of neural network and enhances the training efficiency; on the other hand, the precision of forecasting is also improved markedly. In the process of cases revision, aiming at the forecast errors which are not accepted in abnormal days, some uniform revision formulas are presented for adaptation in this paper. The actual instances show that the proposed model is feasible and promising for STLF.更多还原