Tests with Curved Stopping Boundaries
The stopping rules of Chapter II and III are defined by the crossing of linear boundaries by random walks (or Brownian motion). The linear boundaries arise naturally from sequential probability ratio tests of simple hypotheses against simple alternatives. For problems involving several parameters or composite hypotheses we shall want to consider curved stopping boundaries, which are more difficult to investigate; and only rarely can one obtain exact results even for Brownian motion.
KeywordsBrownian Motion Sequential Probability Ratio Test Generalize Likelihood Ratio Repeated Significance Maximum Sample Size
Unable to display preview. Download preview PDF.