Application of an extremization method to a linear integral of a statistical decision problem