D-最优设计的分部搜索数值构造方法  

Numerical construction algorithm by parted search for D-optimal design

作　　者：乔永辉[1] 李丹东[1] 苏宝库[2] 

机构地区：[1]北京航天控制仪器研究所,北京100039 [2]哈尔滨工业大学控制科学与工程系,哈尔滨150001

出　　处：《中国惯性技术学报》2011年第6期754-759,共6页Journal of Chinese Inertial Technology

基　　金：国家安全重大基础研究项目(973-61334)

摘　　要：为了简化D-最优设计的过程,便于工程应用,提出了一种D-最优设计的数值构造方法——分部搜索数值构造方法。该方法以惯性器件测试为背景,以D-最优性为准则,将传统的D-最优设计分解为谱点寻优和测度寻优两个部分,降低了设计的复杂度,从而减少了运算量,又具有操作简单、思路清楚直观的特点,工程实用性强。提出并证明了相关的定理及推论,从理论上说明了这种构造方法的正确性。以陀螺加速度计误差模型系数辨识试验为例,运用此算法进行了D-最优设计。设计的结果证明了该方法的正确性,同直接搜索法相比,能将最小设计维数降低为原来的50%,提高了设计效率。In order to simplify the process of D-optimal design and facilitate engineering applications,a numerical constructing algorithm is presented,which is a numerical constructing algorithm by dividing the searching into two parts for D-optimal design.Based on the inertial instruments test and the guide rule of D-optimality,the algorithm reduces the design complexity by decomposing the conventional D-optimal design into two parts,i.e.the searching of optimal spectrum points and the searching of optimal weights,and it not only significantly reduces the amount of computation,but also has the characteristics of simple operation and strong practicability.The paper has also proposed and proved the relevant theorem and inferences,which show the correctness of this constructing algorithm theoretically.As an example,this algorithm is used in D-optimal design for PIGA test to identify the error model coefficients.The design results show that the algorithm is correct,and can decrease the minimum design dimension by 50% and improve the design efficiency.

关 键 词：D-最优设计 数值构造方法 分部搜索方法 测试技术 

分 类 号：U666.1[交通运输工程—船舶及航道工程]