Page 306 - Schaum's Outline of Theory and Problems of Advanced Calculus
P. 306

CHAP. 11]                         INFINITE SERIES                               297


                              To obtain the Lagrange form of the remainder R n ,consider the form
                                                               1                K
                                                      0           00     2            n
                                                               2!               n!
                                           f ðxÞ¼ f ðcÞþ f ðcÞðx   cÞþ  f ðcÞðx   cÞ þ     þ  ðx   cÞ
                                                             K
                                                                   n
                              This is the Taylor polynomial P n 1 ðxÞ plus  ðx   cÞ : Also, it could be looked upon as P n except that
                                                             n!
                           in the last term, f  ðnÞ ðcÞ is replaced by a number K such that for fixed c and x the representation of f ðxÞ is
                           exact.  Now define a new function
                                                              n 1         j       n
                                                              X
                                                                 f  ð jÞ  ðx   tÞ  Kðx   tÞ
                                                                       j!      n!
                                                ðtÞ¼ f ðtÞ  f ðxÞþ  ðtÞ    þ
                                                              j¼1
                              The function   satisfies the hypothesis of Rolle’s Theorem in that  ðcÞ¼  ðxÞ¼ 0, the function is
                           continuous on the interval bound by c and x, and   exists at each point of the interval. Therefore, there
                                                                 0
                           exists   in the interval such that   ð Þ¼ 0. We proceed to compute   and set it equal to zero.
                                                                              0
                                                    0
                                                   n 1          j  n 1        j 1       n 1
                                                   X               X
                                                     f               f
                                         0    0       ð jþ1Þ  ðx   tÞ  ð jÞ  ðx   tÞ  Kðx   tÞ
                                                             j!           ð j   1Þ!  ðn   1Þ!
                                          ðtÞ¼ f ðtÞþ    ðtÞ            ðtÞ
                                                   j¼1             j¼1
                           This reduces to
                                                       f  ðnÞ          K
                                                  0       ðtÞ    n 1           n 1
                                                       ðn   1Þ!      ðn   1Þ!
                                                   ðtÞ¼     ðx   tÞ        ðx   tÞ
                           According to hypothesis: for each n there is   n such that
                                                               ð  n Þ¼ 0
                           Thus
                                                             K ¼ f  ðnÞ  ð  n Þ
                           and the Lagrange remainder is
                                                               f  ðnÞ ð  n Þ  n
                                                                n!
                                                         R n 1 ¼    ðx   cÞ
                           or equivalently
                                                           1               nþ1
                                                               f  ðnþ1Þ
                                                     R n ¼        ð  nþ1 Þðx   cÞ
                                                         ðn þ 1Þ!
                              The Cauchy form of the remainder follows immediately by applying the mean value theorem for
                           integrals.  (See Page 274.)

                     11.53. Extend Taylor’s theorem to functions of two variables x and y.
                              Define FðtÞ¼ f ðx 0 þ ht; y 0 þ ktÞ,then applying Taylor’s theorem for one variable (about t ¼ 0Þ
                                                  1           1           1
                                             0      00  2        ðnÞ  n      F  ðnþ1Þ ð Þt nþ1 ;  0 <  < t
                                                 2!           n!        ðn þ 1Þ!
                                  FðtÞ¼ Fð0Þþ F ð0Þþ  F ð0Þt þ     þ  F ð0Þt þ
                           Now let t ¼ 1
                                                                 1          1         1
                                                            0      00          ðnÞ        F ðnþ1Þ
                                                                2!          n!      ðn þ 1Þ!
                                  Fð1Þ¼ f ðx 0 þ h; y 0 þ kÞ¼ Fð0Þþ F ð0Þþ  F ð0Þþ     þ  F ð0Þþ  ð Þ
                              When the derivatives F ðtÞ; ... ; F ðtÞ; F  ðnþ1Þ ð Þ are computed and substituted into the previous expres-
                                                      ðnÞ
                                               0
                           sion, the two variable version of Taylor’s formula results. (See Page 277, where this form and notational
                           details can be found.)
                                   2
                     11.54. Expand x þ 3y   2in powers of x   1 and y þ 2.  Use Taylor’s formula with h ¼ x   x 0 ,
                           k ¼ y   y 0 , where x 0 ¼ 1 and y 0 ¼ 2.
   301   302   303   304   305   306   307   308   309   310   311