• COMPROMISE ALLOCATION FOR TWO-STAGE CLUSTER SAMPLING USING DYNAMIC PROGRAMMING TECHNIQUE

Abstract

In multivariate population where more than one (say ) characteristics are defined on each unit of the population there is no single optimality criterion that can lead us to an optimal sampling design. This problem can be handled efficiently by using a compromise criterion that is optimal for all characteristics in some sense. In the present paper a “compromise allocation” is worked out for the two-stage cluster sampling. The problem is formulated as a Non Linear Programming Problem (NLPP). The NLPP is then solved using Dynamic Programming Technique. A numerical example is also presented to illustrate computational details.