19.3 Properties of the eigenvectors 539
so that, whenλ
k1 ≠ 1 λ
l,
x
lTx
k1 = 10 (19.19)
and the vectors are orthogonal.
EXAMPLE 19.6 For the eigenvectors of the symmetric matrix in Example 19.4(ii),
0 Exercises 16 –18
Property 3.For a symmetric matrix, the eigenvectors corresponding to the same
eigenvalue are either orthogonal or can be made so.
EXAMPLE 19.7The eigenvectorsx
2andx
3of Example 19.4(ii), belonging to the
degenerate eigenvalueλ 1 = 11 , are notorthogonal. Thus
If the independent eigenvectors x
kand x
lcorrespond to the same eigenvalue
λ
k1 = 1 λ
l1 = 1 λ, then
Ax
k1 = 1 λx
kand Ax
l1 = 1 λx
l(19.20)
so that every linear combination ofx
kandx
lis also an eigenvector of Awith the same
eigenvalue. Thus, if
x 1 = 1 ax
k1 + 1 bx
l(19.21)
then
Ax 1 = 1 A(ax
k1 + 1 bx
l) 1 = 1 a(Ax
k) 1 + 1 b(Ax
l)
= 1 a(λx
k) 1 + 1 b(λx
l) 1 = 1 λ(ax
k1 + 1 bx
l) 1 = 1 λx
(19.22)
xx
2 3 23 2311 2
1
2
3
126
T=−
−
cc() ()=++cc )≠ 0
xx
13 13 12111
1
2
3
123
T=
−
cc() ()=+−cc == 0
xx
1 2 12 12111
1
1
2
112
T=
−
cc() ()=+−cc == 0