ASYMPTOTIC PROPERTIES OF A TWO-STAGE PROCEDURE AND THEIR APPLICATIONS
Several optimal fixed sample sizes which arise from sequential problems may be written in the form where q and h are known positive numbers, but q is the unknown and positive nuisance parameter. Mukhopadhyay and Duggan [14] proposed a general two-stage procedure and showed its asymptotic properties as under the condition where is known positive value. We investigate asymptotic properties of their two-stage procedure and show explicitly the relation between the expectation of a given function of the procedure and the known value Our asymptotic expansions are applied to four sequential problems.
sequential estimation, chi-square distribution, bounded risk, fixed- width confidence interval, selection and ranking.