Page 468 - DMTH504_DIFFERENTIAL_AND_INTEGRAL_EQUATION
P. 468

Unit 31: Fredholm Equations with Poincere Goursat Kernels




          Such a Kernel has been mentioned in the case of Volterra integral equations and are  called  Notes
          degenerate Kernels or Poincere Goursat Kernels i.e. P-G Kernels. In equation (2) the functions
          g (x) and h (u), for i = 1, 2, ...n, are L -class. We will see that in this case the Fredholm integral
           i       i                   2
          equation can be reduced to an algebraic system of a linear equation in n unknown. Here g (x) and
                                                                                 i
          h (x) i = 1, 2, ...n are independent in the basic interval (0, 1).
           i
          Substituting (2) in one we have
                           n
                         1
                                  u
                                h
                   x
                  y ( )     g  i ( ) ( ) ( )du  f  ( )
                                             x
                                      u
                              x
                                    y
                                 i
                         0
                          i  1
                        n
                                 1
                   x
          or      y ( )  ( ) g i ( )  h i ( ) ( )du  f  ( )
                                    u
                                              x
                                       u
                                     y
                              x
                                0
                       i  1
          If we put
                      1
                       h  ( ) ( )du                             (x = 1, 2, ...n)   ...(3)
                         u
                             u
                           g
                   k    k
                      0
          then equation (1) becomes
                              n
                   x
                         x
                                 g
                                   x
                  y ( )  f ( )  i i ( )                                            ...(4)
                             i  1
          From equation (4) it is already seen that the difference y(x)  f(x) must necessarily coincide with
          a suitable linear combination of the functions g (x). Now multiply equation (4) by h (x) i = 1, 2,...n
                                               i                           k
          and integrate between 0 and 1 we have
                                             n
                   1            1                 1
                                   x
                                                         x
                                                     x
                      x
                                     f
                    h k ( ) ( )dx  h k ( ) ( )dx  i  g i ( ) ( )dx
                        y
                                      x
                         x
                                                      h
                                                       k
                   0           0                 0
                                             i  1
          Defining
                       h
                   g i ( ) ( )dx  a ik
                         x
                     x
                        k
                                                                                   ...(5)
                   h i ( ) ( )dx  b i
                        x
                     x
                       f
          We have
                       n
                         a    b                                                    ...(6)
                   k      ik i  k
                       i  1
          We thus see that the unknowns  , ...  must satisfy the following system of linear equations
                                     1   n
                  (1  a 11 )  1  a 12 2  a 13 3 ......  a 1n n  b 1
                    a 21 1  (1  a 22 )  2  a 23 3 ....  a 2n n  b 2
                                                                ...(7)
                    a n  1 1  a n  2 2 ...... (1  a nn )  n  b n
          To each set of solution  ° , °  ... °  of this system there corresponds a solution of equation (1)
                               1  2  n
          given by (4). Now the solution of equation exist if the determinant formed by the coefficients
                                                                                      i
          in equation (7) defined by
                        1  a 11  a 12  a 13  ......  a 1n
                          a  (1  a  ) ...............  a
                           21    22           2n
                  D ( )   a 31  ........................................
                                                                                   ...(8)
                        ...................................................
                          a n  1  a n 2  ..............(1  a nn )
                                           LOVELY PROFESSIONAL UNIVERSITY                                   461
   463   464   465   466   467   468   469   470   471   472   473