Page 213 - DMTH504_DIFFERENTIAL_AND_INTEGRAL_EQUATION
P. 213

Differential and Integral Equation




                    Notes          are called the expansion coefficients of f(x) with respect to the given orthonormal system.
                                                            2
                                                      n
                                   Obviously      f    C     dx    0                                       ...(ii)
                                                      1

                                   By writing out the square and integrating term by term, we get
                                                                            n            n
                                                                      2
                                                              0      f dx  2  C   . f  dx  C  2
                                                                             1            1

                                                                           n      n
                                   or                         0    (Nf  ) 2  2  C  2  C  2    [Nf means norm of f]
                                                                            1      1

                                                                          n
                                   or                         0    (Nf  ) 2  C  2
                                                                           1
                                                           n
                                   or                       C  2   (Nf) 2                                 ...(iii)
                                                           1
                                   Since the number on right is Independent of n, it follows that

                                                           n
                                                            C  2  < (Nf) 2.
                                                           1
                                   This fundamental inequality is known as  Bessel’s inequality and is true for every orthonormal
                                   system. It proves that the sum of the squares of the expansion coefficients always converges.

                                   For systems of functions with complex values the corresponding relation is
                                                          n
                                                              2        2
                                                                            f
                                                                          f
                                                            C      (Nf  )  ( , )                          ...(iv)
                                                           1
                                                                           f
                                   holds, where C  is the expansion coefficient  C  ( ,  ) .
                                   This relation may be obtained from the inequality
                                                            2
                                                      n                   n
                                                  x
                                                 f  ( )  C   dx = (Nf  ) 2  C  2  0
                                                       1                   1
                                   The significance of the integral in (ii) is that it occurs in the problem of approximating the given
                                                                 n
                                   function f(x) by a linear combination    with   as constant coefficient and fixed n, in such
                                                                  1
                                   a way that the mean square error
                                                                                2
                                                                         n
                                                              M =     f         dx
                                                                          1

                                   is as small as possible.
                                   An approximation of this type is known as an approximation by the method of least squares, or
                                   an approximation in the mean.




          206                               LOVELY PROFESSIONAL UNIVERSITY
   208   209   210   211   212   213   214   215   216   217   218