OPTIMUM INPUT SIGNALS FOR PARAMETER-ESTIMATION IN SYSTEMS DESCRIBED BY LINEAR INTEGRAL-EQUATIONS

作     者：RAFAJLOWICZ, E 

作者机构：WROCLAW TECH UNIVWYBRZEZE WYSPIANSKIEGO 27PL-50370 WROCLAWPOLAND 

出 版 物：《COMPUTATIONAL STATISTICS & DATA ANALYSIS》 (计算统计学与数据分析)

年 卷 期：1990年第9卷第1期

页      面：11-19页

核心收录：

学科分类：08[工学] 0714[理学-统计学（可授理学、经济学学位）] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

主　　题：Experiment design D-Optimality integral equation description Information matrix Numerical algorithm 

摘      要：The paper deals with the problem of optimum experimental design for estimation of parameters, which appear in the kernel of an integral equation describing a physical process. The autocorrelation function of a spatial random field acting as an input (stimulus) is the decision variable in the considered problem. Under integral constraints imposed on decision variables results parallel to the optimum experiment design theory of Kiefer and Wolfowitz are presented. In particular, necessary and sufficient optimality conditions are derived and a computational algorithm is proposed.