Let X1; X2; ... be a discrete-time stochastic process with a distribution Py, y 2 Y, where
Y is an open subset of the real line. We consider the problem of testing a simple
hypothesis H0: y ¼ y0 vs. a composite alternative H1: y4y0, where y0 2 Y is some fixed
point. The main goal of this article is to characterize the structure of locally most
powerful sequential tests in this problem.
For any sequential test ðc;fÞ with a (randomized) stopping rule c and a
(randomized) decision rule f let aðc;fÞ be the type I error probability, _
b0ðc;fÞ the
derivative, at y ¼ y0, of the power function, and N ðcÞ an average sample number of the
test ðc;fÞ. Then we are concerned with the problem of maximizing _
b0ðc;fÞ in the class
of all sequential tests such that
aðc;fÞra and N ðcÞrN ;
where a 2 ½0; 1 and NZ1 are some restrictions. It is supposed that N ðcÞ is calculated
under some fixed (not necessarily coinciding with one of Py) distribution of the process
X1; X2; ... .
The structure of optimal sequential tests is characterized.