Page 479 - DMTH504_DIFFERENTIAL_AND_INTEGRAL_EQUATION
P. 479

Differential and Integral Equation




                    Notes          For this reason, the Kernels  (s, t) and   (s, t) are called the resolvent Kernels of the equation (1)
                                   and (2) respectively.
                                   The foregoing  theorem shows  that    is not an  eigenvalue of either the Kernel  K(s, t) or  its
                                   conjugate Kernel K (s, t),
                                                          K (s, t) = K(t, s)                              ...(15)
                                   The General Case

                                   We shall prove that there exist two sets of linearly independent continuous functions
                                                 s
                                                        s
                                             s
                                            ( ),  ( ), ...,  ( )
                                           1    2      m
                                                                                                          ...(15)
                                          B 1 ( ),  2 ( ), ...,  m ( )
                                             t
                                                 t
                                                        t
                                   defined on the interval [a, b], such that
                                                                2
                                           b  b       m
                                              K ( , )    v ( ) ( ) ds dt  1                               ...(16)
                                                          s
                                                s
                                                  t
                                                              t
                                                             v
                                           a  a
                                                     v  1
                                   To prove this, let   be an arbitrary positive number. Then we divide the interval (a, b) into a finite
                                   number of sub-intervals I , I , ..., I , such that
                                                       1  2   n
                                                       t
                                                   K
                                                     s
                                                            s
                                          sup     | ( , )  K ( , t  | )
                                             a  s  v
                                   for any pair of points  t , t  in each  I .  This is  possible, because of the uniform continuity  of
                                                                 v
                                   K(s, t) on the domain a    s    b, a    t    b. Let t  be an inner point of I . Let I  be an interval
                                                                          v                  v     v
                                   contained in the interior of I  and containing the point t . Then we define  (t) as follows:
                                                         v                      v              v
                                                 0 outside of I v
                                            ( )
                                             t
                                           v
                                                 1    on I  v
                                   such that the function  (t) is continuous and 0      (t)    1 on the interval [a, b]. We now set
                                                      v                      v
                                           s
                                     s
                                     ( ) K ( , ) and
                                            t
                                    r
                                                         n
                                                             s
                                                   s
                                                     t
                                                                 t
                                            s
                                          N ( , )  K ( , )  v ( ) ( )
                                              t
                                                                v
                                                        v  1
                                   Then we see that
                                                       t
                                                            s
                                                     s
                                             s
                                               t
                                          | ( , )| | ( , )  K ( , t v  )|
                                           N
                                                   K
                                   for t in I , and
                                          v
                                             s
                                                                  t
                                           N
                                                     s
                                                       t
                                                            s
                                               t
                                          | ( , )| | ( , )  K ( , t v  ) ( )|  2 M
                                                   K
                                                                v
                                   for t in I    I  where
                                         v   v
                                          M =  sup  a  s  b  a  t  b  | ( ,  | ) t
                                                                 K
                                                                   s
                                   Since   and the sum of lengths of I    I  are both arbitrary, we can choose the values of them so
                                                              v   v
                                   small that
                                                                2
                                                      n
                                           b  b
                                                              t
                                              K ( , )    v ( ) ( ) ds dt  1
                                                          s
                                                  t
                                                s
                                                             v
                                           a  a
                                                     v  1
                                   Clearly, the function  (t),  (t), ...,   (t) are linearly independent. Hence, if   (s),  (s), ...,   (s)
                                                     1   2      n                                1   2      n
                                   are linearly independent, then our proof is completed. If otherwise, say,   (s) is written as a
                                                                                                n
                                   linear combination of   (s),   (s), ...,  – (s), then
                                                      1   2      n 1
          472                               LOVELY PROFESSIONAL UNIVERSITY
   474   475   476   477   478   479   480   481   482   483   484