误差环境中参数识辨前测量信息的熵描述  

Entropy description of measured information with error before parameter identification

作　　者：李静[1] 丁海洋 任学尧 Li Jing;Ding Haiyang;Ren Xueyao(Xijing University,Xi'an 710000,China;College of Information & Communication,National University of Defense Technology,Xi' an 710106,China;Air Force Engineering University,Xi' an 710043,China)

机构地区：[1]西京学院,西安710000 [2]国防科技大学信息通信学院,西安710106 [3]空军工程大学,西安710043

出　　处：《计算机应用研究》2019年第7期1988-1990,共3页Application Research of Computers

基　　金：国家自然科学基金青年基金资助项目(61301135)

摘　　要：在逆问题的参数识辨中,测量信息包含的信息量直接影响参数的重构精度,如何度量测量信息中的信息量,对选取参数识辨所需的测量点个数具有重要的指导作用。根据贝叶斯方法,将体现先验信息的先验概率和反映测量信息的似然概率融合得到待识辨参数样本出现的概率,结合信息熵和最大熵计算评价因子,将评价因子用于描述参数识辨前带有误差的测量信息所包含的信息量。计算实例表明,该计算方法可以有效地描述误差环境中的测量信息,对实际应用中测量信息的选取具有重要的指导意义。In the parameter identification of inverse problems,it was well-known that the information quantity contained in measurement information influences the reconstruction precision of parameter directly.How to describe the information quantity in measurement information plays an important role in the selection of the number of measurement points.According to the Bayesian method,it calculated the probability of the parameter sample to be identified by combining the priori probability reflecting prior information and the likelihood probability reflecting measurement information.It calculated the evaluation factors by the maximum entropy and information entropy of the parameter.This paper introduced the information entropy to describe the information quantity of measurement information with error before parameter identification.Numerical tests show that this proposed method can describe the measured information with error effectively and determine the measured information in practice.

关 键 词：参数识辨 信息熵 贝叶斯方法 误差 评价因子 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]