Page 274 - DMTH502_LINEAR_ALGEBRA
P. 274

Linear Algebra




                    Notes          This function f  is a linear functional on V, because by its very definition, (  ) is linear as a function
                                   of  . If V is finite-dimensional, every linear functional on V arises in this way from some  .
                                   Theorem 1: Let V be a finite-dimensional inner product space, and f a linear functional on V. Then
                                   there exists a unique vector  in V such that f( ) = (  ) for all  in V.
                                   Proof: Let { ,  , ....,   } be an orthonormal basis for V. Put
                                             1  2    n
                                                                    n
                                                                 =     ( f  j  )  j                        ...(1)
                                                                    j  1

                                   and let f  be the linear functional defined by
                                                            f ( ) = (  ).
                                   Then


                                                           f ( ) =   k    ( f  j )  j  ( f  k  )
                                                              k
                                                                        j
                                   Since this is true for each  , is follows that f = f . Now suppose   is a vector in V such that (  )
                                                        k
                                   = (  ) for all  . Then (   ) = 0 and   . Thus there is exactly one vector   determining
                                   the linear functional f in the stated manner.

                                   The proof of  this theorem can be reworded slightly, in terms of the representation of  linear
                                   functionals in a basis. If we choose on orthonormal basis {  , ....,   ) for V, the inner product of
                                                                                  1     n
                                    = x   + ... + x   and   = y   + ...+ y   will be
                                       1  1    n  n       1  1    n  n
                                                           (   ) = x y  ... x y  .
                                                                    1 1
                                                                            n n
                                   If f is any linear functional on V, then f has the form
                                                             f( ) = c x  + ... +c x
                                                                    1 1    n n
                                   for some fixed scalars c , ...., c  determined by the basis. Of course c = f ( ). If we wish to find a
                                                     1    n                             j    j
                                   vector   in V such that (  ) = f ( ) for all  , then clearly the coordinates y  of   must satisfy
                                                                                                 j
                                   y  i  c j  or  y i  ( f  j ). Accordingly,

                                                                 =  ( f  1 )  1  ...  ( f  n )  n
                                   is the desired vector.
                                   Some further comments are in order. The proof of Theorem 1 that we have given is admirably
                                   brief, but it fails to emphasize the essential geometric fact that   lies in the orthogonal complement
                                   of the null space of f. Let W be the null space of f. Then V = W + W , and f is completely determined
                                   by its values on W . In fact, if P is the orthogonal projection of V on W , then
                                                             f( ) = f (P )

                                   for all   in V. Suppose f  0. Then f is of rank 1 and dim (W ) = 1. If   is any non-zero vector in W ,
                                   it follows that
                                                                   (   )
                                                             P   =     2

                                   for all   in V. Thus
                                                                         f  ( )
                                                             f( ) = (  ).
                                                                            2



          268                               LOVELY PROFESSIONAL UNIVERSITY
   269   270   271   272   273   274   275   276   277   278   279