International Journal of Mathematics and Mathematical Sciences
Volume 18 (1995), Issue 1, Pages 147-150
doi:10.1155/S0161171295000196

Asymptotic optimality of sequential designs for estimation

Abstract

This paper is concerned with the problem of allocating a fixed number of trials between K independent populations from the exponential family, in order to estimate a linear combination of the means with squared error loss. Introducing independent conjugate priors, a batch sequential procedure is proposed and compared with the optimal.