基于加权平均一阶发散度的混沌序列预测法  

Forecasting method of chaotic sequence based on weighted average first-order divergence degree

作　　者：岳毅宏[1] 韩文秀[1] 王健[1] 

机构地区：[1]天津大学管理学院,天津300072

出　　处：《系统工程与电子技术》2004年第5期602-604,共3页Systems Engineering and Electronics

基　　金：国家自然科学基金资助课题(79970043)

摘　　要：深入分析了基于最大Lyapunov指数预测法产生误差的根源。在此基础上定义了一个新的变量:加权平均一阶发散度,并基于该变量提出了一种新的混沌序列预测方法。首先从理论上对该方法的基本原理进行了系统论述,并指出了加权平均一阶发散度所具有的一些显著特点。然后总结了所提预测方法的算法过程。最后将新方法应用于电力系统的负荷预测中,得到了理想的预测结果。通过分析和比较,验证了其有效性。The error cause of the forecasting method based on maximal Lyapunov exponent is thoroughly analyzed. On this basis, a new variable called weighted average first-order divergence degree is defined. Based on this variable, a novel forecasting method of chaotic sequence is proposed. Firstly, its principle is demonstrated systematically from the theoretical aspect, and some remarkable characteristics of weighted average first-order divergence degree are pointed out. Then, arithmetic procedure of the proposed forecasting method is summarized. In the end, this method is applied to the forecasting of short-term load, and the results are ideal. The analytic and comparison results prove its validity.

关 键 词：LYAPUNOV指数 混沌序列 一阶发散度 电力负荷预测 

分 类 号：O415.5[理学—理论物理]