Page 193 - DMTH504_DIFFERENTIAL_AND_INTEGRAL_EQUATION
P. 193

Differential and Integral Equation




                    Notes          where a superscript asterisk denotes the complex conjugate. This means that the inner product
                                   has the properties
                                   (i)    ,    =   ,
                                         1  2   2  1
                                                  *
                                   (ii)  a  , a   = a a  ,
                                         1 1  2 2  1  2  1  2
                                   (iii)  ,   +    =   ,    +   ,   ,    +  ,    =   ,    +   ,
                                         1  2  3   1  2    1  3  1  2  3   1  3    2  3
                                                  2
                                   (iv)  ,   =   | |  dx   0, with equality if and only if  (x)   0 in I.
                                              I
                                   Note that this reduces to the definition of a real inner product if   and   are real. If   ,    = 0
                                                                                       1     2          1  2
                                   with      0 and      0, we say that   and   are orthogonal.
                                        1        2               1    2
                                                  2
                                   Let y (x), y (x)   C  [a, b] be twice-differentiable complex-valued functions. By integrating by
                                       1   2
                                   parts, it is straightforward to show that
                                                                                   
                                                                 x
                                           y Sy   Sy y    p ( ){ ( )(y  ( ))  y  ( ) ( )}                  ...(7)
                                                              y
                                                            x
                                                                            x
                                                                              y
                                                                                x
                                                                     x
                                            2  1    2  1       1    2      1   2   
                                   which is known as Green’s formula. The inner products are defined over a sub-interval [ ,  ]
                                                                    +
                                   (a, b), so that we can take the limits     a  and    b  when the endpoints are singular, and the
                                   Sturm-Liouville operator, S, is given by (3). Now if x = a is a regular endpoint and the function
                                   y  and y  satisfy a separated boundary condition at a, then
                                    1    2
                                                              *
                                                     a
                                            a
                                              y
                                                           a
                                                             y
                                                               a
                                                a
                                          p ( ){ ( )(y  * 2 ( ))  y 1 ( ) ( )}  0.                         ...(8)
                                               1
                                                              2
                                   If a is a finite singular endpoint and the functions  y  and  y  satisfy the Friedrich’s boundary
                                                                              1    2
                                   condition at a,
                                                                  *
                                                       *
                                                  y
                                                               x
                                                                 y
                                                                   x
                                                     x
                                                      y
                                                         x
                                                x
                                               p
                                           lim [ ( ){ ( ) ( ))  y 1 ( ) ( )}] 0
                                                   1
                                                                  2
                                                       2
                                          x  a                                                             ...(9)
                                   Similar results hold at x = b.
                                   We can now derive several results concerning the eigenvalues and eigenfunctions of a Sturm-
                                   Liouville boundary value  problem.
                                   Theorem 1: The eigenvalues of a Sturm-Liouville boundary value problem are real.
                                                                   x
                                                                                   x
                                                                             x
                                                                                 y
                                                                               
                                                           y  * ( , )Sy ( , )  Sy * ( , ), ( , )
                                                                                     
                                                             x
                                                               
                                                                     
                                                                                 
                                                                   *
                                                            y
                                                                
                                                                      
                                                              x
                                                                                  x
                                                                              
                                                        [ ( ){ ( , )( ( , ))  y ( , ) ( , )}]  b  0
                                                                                    
                                                                  y
                                                                                y
                                                                             x
                                                          x
                                                                    x
                                                         p
                                                                                       a
                                                                            *
                                   Proof: If we substitute y (x) = y(x,  ) and  y (x) = y (x,  ) into Green’s formula over the entire
                                                      1               2
                                                                 x
                                                          x
                                                                               y
                                                                                x
                                   interval, [a, b], we have  y  *( , ), Sy ( , )  Sy *( , ), ( , )
                                                                           x
                                                                                        b
                                                                     x
                                                                             x
                                                              x
                                                          x
                                                        p ( ) y ( , )( *( , )  y  ( , ) *( , )  0
                                                                                   x
                                                                                y
                                                                  y
                                                                                        a
                                   making use of (8) and (9). Now, using the fact that the function y(x,  ) and y*(x,  ) are solutions
                                   of (1) and its complex conjugate, we find that
                                           b         *                  *  b         2
                                              x
                                                    y
                                                 x
                                               y
                                                                               y
                                                      x
                                            r ( ) ( , ) ( , )(    ) dx  (    )  r ( )[ ( , )] dx  0
                                                                                 x
                                                                             x
                                                                                   
                                                        
                                                  
                                           a                              a
                                   Since r(x) > 0 and y(x,  ) is nontrivial, we must have   =  * and hence    . i.e. the eigenvalues
                                   are real.
                                   Theorem 2: If y(x,  ) and  y(x,   )  are eigenfunctions  of the  Sturm-Liouville boundary  value
                                                                                           p
                                   problem, with      ), then these eigenfunctions are orthogonal over C [a, b] with respect to the
                                   weighing function r(x), so that
                                           b
                                                    y
                                               y
                                                 x
                                              x
                                            r ( ) ( , ) ( , ) dx  0                                       ...(10)
                                                      x
                                                        
                                                   
                                           a
          186                               LOVELY PROFESSIONAL UNIVERSITY
   188   189   190   191   192   193   194   195   196   197   198