Page 217 - Linear Algebra Done Right
P. 217

Now we can prove a result that was promised in the last chapter,
                      where we proved the analogous theorem (8.20) for operators on com-
                      plex vector spaces.  The Characteristic Polynomial                                   207
                      9.20   Cayley-Hamilton Theorem: Suppose V is a real vector space
                      and T ∈L(V). Let q denote the characteristic polynomial of T. Then
                      q(T) = 0.

                         Proof: Choose a basis of V with respect to which T has a block   This proof uses the
                      upper-triangular matrix of the form 9.18, where each A j is a 1-by-1  same ideas as the proof
                      matrix or a 2-by-2 matrix with no eigenvalues. Suppose U j is the one- or  of the analogous result
                      two-dimensional subspace spanned by the basis vectors corresponding  on complex vector
                      to A j . Define q j as in 9.19. To prove that q(T) = 0, we need only show  spaces (8.20).
                                  = 0 for j = 1,...,m. To do this, it suffices to show that
                      that q(T)| U j
                      9.21                   q 1 (T)...q j (T)| U j  = 0

                      for j = 1,...,m.
                         We will prove 9.21 by induction on j. To get started, suppose that
                                                                          = 0 (obvious if
                      j = 1. Because M(T) is given by 9.18, we have q 1 (T)| U 1
                      dim U 1 = 1; from 9.7(a) if dim U 1 = 2), giving 9.21 when j = 1.
                         Now suppose that 1 <j ≤ n and that


                                           0 = q 1 (T)| U 1
                                           0 = q 1 (T)q 2 (T)| U 2
                                            . . .

                                                                  .
                                           0 = q 1 (T)...q j−1 (T)| U j−1
                      If v ∈ U j , then from 9.18 we see that
                                             q j (T)v = u + q j (S)v,

                      where u ∈ U 1 + ··· + U j−1 and S ∈L(U j ) has characteristic poly-
                      nomial q j . Because q j (S) = 0 (obvious if dim U j = 1; from 9.7(a) if
                      dim U j = 2), the equation above shows that


                                           q j (T)v ∈ U 1 + ··· + U j−1
                      whenever v ∈ U j . Thus, by our induction hypothesis, q 1 (T)...q j−1 (T)
                      applied to q j (T)v gives 0 whenever v ∈ U j . In other words, 9.21 holds,
                      completing the proof.
   212   213   214   215   216   217   218   219   220   221   222