改进范例推理在短期负荷预测中的应用

Improved Case-based Reasoning for Short-term Load Forecasting

范例推理是人工智能领域中一种新兴的推理方法,基于范例的推理就是充分利用以前的经验进行推理,求解新问题的过程。文中针对短期负荷预测具有明显周期性的特点,提出了基于范例推理的短期负荷预测方法。首先说明了范例的表示、组织方法,然后选择合适的量度方法进行范例的检索、匹配,最后将得到的最相似范例集进行重用、修正,得到最终预测结果。在范例表示中,使用粗糙集方法为范例属性特征的选择及权重的确定问题提供了一个合理的方法。经实例验证和比较,基于范例推理的负荷预测方法在提高预测精度方面具有明显的优越性。

Short-term load forecasting (STLF) plays a key role in power system operation and planning. As STLF has obvious periodicity, this paper presents a hybrid approach combining case-based reasoning (CBR) and rough set for STLF problems. CBR consists of the steps of case representation, indexing, retrieval, and adaptation, and the crucial concept in CBR involves the use of the already existing knowledge about objects or situations to predict the relevant aspects of similar objects. The rough set theory is used to reduce the initial information table in order to obtain the refined attributes that constitute a case. This method provides a new way of selecting the relevant feature subset and feature weights. The testing results concerning an actual power system show that the model proposed is feasible and promising for load forecasting.