Page 222 - DMTH502_LINEAR_ALGEBRA
P. 222

Linear Algebra




                    Notes          which shows that  s   2. Now it  makes sense  to  ask whether  or not  p  = g . From the  two
                                                                                              2   2
                                   decompositions of V, we obtain two decompositions of the subspace p V:
                                                                                            2
                                          p V  p W 0  ( Z p 2  1 ; )
                                                           T
                                                2
                                           2
                                                                                                           ...(12)
                                                                        T
                                          p V  p W    ( Z p  ; )  ... Z (p  ; ).
                                                           T
                                           2    2  0    2 1          2 s
                                   We have made use of facts (1) and (2) above and we have used the fact that  p 2  i  0, i  2.  Since
                                                                                                T
                                                                                   T
                                                                                          Z
                                   we know that p  = g , fact (3)  above  tells us that  (Z p 2  1 ; )  and  (p 2 1 ; )   have the  same
                                                1   1
                                   dimension. Hence, it is apparent from (12) that
                                                   T
                                              Z
                                          dim (p   ; ) 0,       i  2.
                                                2 i
                                   We conclude that p   = 0 and g  divides p . The argument can be reversed to show that p  divides
                                                  2 2      2       2                                    2
                                   g . Therefore p  = g .
                                    2         2   2
                                   Corollary: If T is a linear operator on a finite-dimensional vector space, then every T-admissible
                                   subspace has a complementary subspace which is also invariant under T.
                                   Proof: Let W  be an admissible subspace of V. If W  = V, the complement we seek is {0}. If W  is
                                             0                             0                                 0
                                   proper, apply Theorem 1 and let
                                                               T
                                                   T
                                          W 0   ( Z  1 ; )  ...  ( Z  r  ; ).
                                   Then  W  is invariant under T and V = W    W .
                                         0                          0    0
                                   Corollary: Let T be a linear operator on a finite-dimensional vector space V.
                                   (a)  There exists a vector  in V such that the T-annihilator of  is the minimal polynomial for T.
                                   (b)  T has a cyclic vector if and only if the characteristic and minimal polynomials for T are
                                       identical.
                                   Proof: If V = {0}, the results are trivially true. If V  {0}, let
                                                           T
                                                T
                                        V   ( Z  1 ; ) ... Z (  r  ; )                                     ...(13)
                                   where the T-annihilators  p 1 ,....,p  are such that  p k  1  divides  p k  , 1  k  r  1.  As we noted in the
                                                             r
                                   proof of Theorem 1, it follows easily that p  is the minimal polynomial for T, i.e., the T-conductor
                                                                    1
                                   of V into {0}. We have proved (a).
                                   We saw in unit 19 that, if T has a cyclic vector, the minimal polynomial for T coincides with the
                                   characteristic polynomial. The content of (b) is in the converse. Choose any   as in (a). If the
                                   degree of the minimal polynomial is dim V, then V = Z ( ; T).
                                   Theorem 2 (Generalized Cayley-Hamilton Theorem): Let  T  be a linear operator on a finite-
                                   dimensional vector space V. Let p and f be the minimal and characteristic polynomials for  T,
                                   respectively.

                                   1.  p divides f.
                                   2.  p and f have the same prime factors, except for multiplication.
                                   3.  If

                                          p  f  1 T  .... f  k T                                           ...(14)
                                              1   1
                                   is the prime factorization of p, then
                                          f  f  1 1 d  ....f k  k d                                        ...(15)

                                                          n
                                   where d  is the nullity of f (T)  divided by the degree of f .
                                         i             i                        i

          216                               LOVELY PROFESSIONAL UNIVERSITY
   217   218   219   220   221   222   223   224   225   226   227