Filters
A study of the efficiency of chemical balance weighing designs
Abstract Some elements concerning the problem of determining chemical balance designs are considered. A method is presented for determining the best design in classes where no D-optimal design exists. In this case, we obtain a design that is highly D-efficient. The efficiency of such a design is presented as a function of the number of parameters and measurements, and its properties are studied. Charts showing the efficiency function for selected classes of designs are presented.
D-optimal chemical balance weighing designs with positively correlated errors: part I
Summary In this paper we consider issues related to the determination of chemical balance weighing designs satisfying the criterion of D-optimality. This problem is considered under the assumption that the measurements are equally positively correlated and they have the same variances. We answer the question of how to add four runs to a regular D-optimal chemical balance weighing design in order to obtain a highly D-efficient design. The problem formulated in this way leads to an extension of results contained in a previous paper in Biometrical Letters.
D-optimal chemical balance weighing designs with positively correlated errors: part II
Summary Some questions related to the problem of determining a chemical balance weighing design satisfying the criterion of D-optimality are considered. The theory is based on the assumption that the measurements are equally positively correlated and have the same variances. In this paper, we present a method of adding three measurements to a regular D-optimal chemical balance weighing designs to obtain a highly D-efficient chemical balance weighing design. The problem formulated in this way is an extension of results contained in a paper previously published in Biometrical Letters.