综合法优选局域邻近点的供水量预测模型

发布时间：2018-04-18 22:32

本文选题：混沌理论 + 邻近点　；参考：《土木建筑与环境工程》2016年S2期

【摘要】：基于混沌理论的局域法是目前较先进的预测技术,对于非线性、非平稳特征较强的城市日供水系统的预测,通常混沌局域法更适用。混沌局域法的日供水量预测,最重要的就是对历史信息的利用,即对邻近点的选择和利用,但在运用混沌局域法对城市日供水量预测时却鲜有人考虑邻近点选取问题,或仅考虑邻近点个数与邻近点位置中的一个,很少没有人将2种情况综合考虑的研究。据此,运用BIC信息准则和衰减系数相结合对相空间重构进行数据挖掘,优选混沌局域法的邻近点,并用该模型对实际水厂日供水量进行了预测。结果表明,运用该方法能显著提高日供水量预测精度,预测平均绝对误差仅为1.06%,说明该综合法是可行的。
[Abstract]:The local method based on chaos theory is an advanced prediction technique at present. It is more suitable for the prediction of urban daily water supply system with nonlinear and non-stationary characteristics.The most important thing in forecasting daily water supply by chaotic local method is the use of historical information, that is, the selection and utilization of adjacent points. However, few people consider the selection of adjacent points when using chaotic local method to predict daily water supply in cities.Or consider only one of the number of adjacent points and the location of the adjacent points, and few studies have taken the two cases into account.Based on this, the data mining of phase space reconstruction is carried out by using BIC information criterion and attenuation coefficient, and the adjacent points of chaos local method are selected optimally, and the daily water supply quantity of actual water plant is predicted by the model.The results show that the prediction accuracy of daily water supply can be improved significantly by using this method, and the average absolute error of prediction is only 1.06, which shows that the synthetic method is feasible.
【作者单位】：昆明理工大学建筑工程学院;
【基金】：昆明理工大学2016年学生课外学术科技创新基金(2015YB025) 国家自然科学基金(51608242) 云南省人才培养计划项目(14118943)
【分类号】：TU991.31;O415.5

【相似文献】