Percentiles for Cramér-von Mises functionals of Gaussian processes and some applications to Bayesian tests

The main objective for this paper is twofold. We first present a method for the derivation of an arbitrarily exact approximation to the distribution of Cramér-von Mises type functionals of any given Gaussian process X = {X(t): 0[less-than-or-equals, slant]t[less-than-or-equals, slant]1}. Secondly, we demonstrate how this method may be used to easily tabulate previously almost inaccessible distributions such as those of limiting functionals arising in tests for detecting change in regression parameters (cf., e.g., Jandhyala and MacNeill, 1989).