Application of geometric programming to optimum allocation problems in multivariate doubling sampling (Q1091946)

The paper gives a geometric programming approach to solve optimum allocation problems in multivariate double sampling. In section two this optimum allocation problem is introduced, while the next section describes a transformation of this problem into a geometric programming problem. In doing so the constraint are normalized such that lower bounds for the variables and for the value of the objective function can be derived. Furthermore, in this step irrelevant constraints can be eliminated. By this reduction a problem is obtained which can explicitly be solved, however the solution in general is not integer. The author's view is that the non-integer values may be rounded off. The paper ends with a numerical example and a listing of the Fortran program of the proposed method.