المستخلص: |
The logarithms of the treatment parameters in a mixture paired comparison experiments, are approximated by a first-order response surface model in the transformed mathematically independent variables. Asymptotic properties of the likelihood estimators of the regression parameters are presented. A class of partially balanced designs D1, corresponding to null hypotheses about regression parameters is constructed. The optimal design d1 in D1, is the A-, D-, and E- optimal design minimizing, the trace, determinant and largest eigenvalue of a defined covariance matrix is determined
|