Page 243 - Advanced Linear Algebra
P. 243

Chapter 10

            Structure Theory for Normal Operators















            Throughout this chapter, all vector spaces are assumed to be finite-dimensional
            unless otherwise noted. Also, the field   is either   or  .
                                                    s
                                           -
                                                        d
            The Adjoint of a Linear Operator
            The purpose of this chapter is to study the structure of certain special types of
            linear operators on finite-dimensional real and complex inner product spaces. In
            order to define these operators, we introduce another type of adjoint (different
            from the operator adjoint of Chapter 3).

            Theorem 10.1 Let   and  >   be finite-dimensional inner product spaces over  -
                            =

                      B
            and let  ²= Á > ³ . Then there is a unique function     i ¢ > ¦ =  , defined by
            the condition
                                     º#Á $» ~ º#Á    i $»

            for all #=   and $>  . This function is in  ²> Á = ³  and is called the adjoint
                                                 B
            of .

            Proof.  If   exists, then it is unique, for if
                     i



                                     º#Á $» ~ º#Á $»
                                     #
            then º#Á $» ~ º#Á    i  $»  for all   and   and so  ~       i  .
                                          $

                               i
            We seek a linear map   ¢> ¦ =   for which

                                     º#Á    i $» ~ º #Á $»
                                         i
            By way of motivation, the vector   $ , if it exists, looks very much like a linear
                       #
            map  sending    to  º  #  Á    $  »  . The only problem is that     i #   is supposed to be a
            vector, not a linear map. But the Riesz representation theorem tells us that linear
            maps can be represented by vectors.
   238   239   240   241   242   243   244   245   246   247   248