The Minimax Wiener Sequential Testing Problem

Abstract

Consider the sample path of a one-dimensional diffusion for which the diffusion coefficient is given and where the drift may take on one of two values: $\mu_0$ or $\mu_1$. Suppose that the signal-to-noise ratio (defined as the difference between the two possible drifts divided by the diffusion coefficient) is non-constant. Given an initial state for the observed process, we consider a minimax formulation of the Wiener sequential testing problem for detecting the correct drift coefficient as soon as possible and with minimal probabilities of incorrect terminal decisions. We solve the problem in the Bayesian formulation, under any prior probabilities of the process having drift $\mu_0$ or $\mu_1$, when the passage of time is penalized linearly. In the case where the signal-to-noise ratio is assumed constant, we obtain an explicit formula for the least favorable distribution.