Page 304 - Elements of Distribution Theory
P. 304

P1: JZP
            052184472Xc09  CUNY148/Severini  June 2, 2005  12:8





                            290                       Approximation of Integrals

                            Corollary 9.2. Let f denote a continuously differentiable function on [m, ∞] where m is
                            an integer. Assume that
                                                 ∞                ∞

                                                   | f ( j)| < ∞,  | f (x)| dx < ∞
                                                j=m             m
                            and
                                                         ∞


                                                           | f (x)| dx < ∞.
                                                        m
                            Then
                                          ∞           ∞                    ∞
                                                                1

                                             f ( j) =  f (x) dx +  f (m) +  P(x) f (x) dx
                                         j=m        m           2        m
                            where
                                                                      1
                                                        P(x) = x −	x − .
                                                                      2

                            Proof. By Theorem 9.17, for any r = m, m + 1,...,
                                       r          r                           r
                                                            1

                                         f ( j) =  f (x) dx + [ f (m) + f (r)] +  P(x) f (x) dx.  (9.8)
                                      j=m        m          2               m
                            Note that, under the conditions of the corollary,
                                                        r

                                                        |P(x)|| f (x)| dx < ∞
                                                      m
                            and
                                                          lim f (r) = 0.
                                                          r→∞
                            The result now follows from taking limits in (9.8).

                              Using Theorem 9.17, the sum
                                                             r

                                                                f ( j)
                                                            j=m
                            may be approximated by the integral
                                                              r
                                                              f (x) dx
                                                            m
                            with error
                                                 1                  r

                                                  [ f (m) + f (r)] +  P(x) f (x) dx
                                                 2                m
                            or it may be approximated by
                                                     r
                                                               1
                                                       f (x) dx + [ f (m) + f (r)]
                                                    m          2
   299   300   301   302   303   304   305   306   307   308   309