Académique Documents
Professionnel Documents
Culture Documents
Key Issues:
A symmetric yT Ax xT Ay 0 xT y
xT y 0, as wished.
Fall 2018 Prof.Jiang@ECE NYU 162
Canonical Form – First Pass
nn
Consider a real symmetric matrix A ,
with distinct (real, by Fact 1) eigenvalues i i 1 .
n
such that
1 0
O AO diag i .
T
0
n
Fall 2018 Prof.Jiang@ECE NYU 163
Constructive Proof
For each eigenvalue i , take an eigenvector xi ,
which has unit norm, i.e., xi ( xi )T x i 1.
Define a matrix O as:
O x , , x
1 n
nn
x1
T
Then, O
T n n
xn
T
Fall 2018 Prof.Jiang@ECE NYU 164
Constructive Proof (cont’d)
In addition, OT AO diag i .
Indeed , det I P 1 AP det P 1 I A P
det P det I A det P det I A ,
1
noting that det P 1 det P det P 1 P 1.
Fall 2018 Prof.Jiang@ECE NYU 168
Exercise 1
Can the following matrix
1 0
A
1 1
be transformed into the canonical diagonal form
1 0
0 1
Two matrices having the same eigenvalues may not be similar.
Fall 2018 Prof.Jiang@ECE NYU 169
Exercise 2
Can the following matrix
0 1
A
0 0
be transformed into a canonical diagonal form?
AP P Ap i i p i , i 1, 2,...., n
implying that p is an eigenvector for eigenvalue i .
i
n
eigenvectors p i
, i.e., Ap i p .
i i
i 1
Then, P p1 p2 pn is nonsingular and
satisfies (by direct computation) that
P1 AP .
j j
Then, P c c , implying that
1 2
1 j 0
P AP .
0 j
Fall 2018 Prof.Jiang@ECE NYU 177
Diagonalizable Matrix
A matrix is said to be "diagonalizable", if it is similar to
a diagonal matrix.
Are the following statements true or false:
mn
Ax 0, A , x .
n
0
n
a1 0
A
0 a
n
i1 successively,
N
vectors y i
i1 ?
N
n-dimensional vectors x i
x
N
i
. Here is the systematic procedure.
i 1
First ,
y1 : x1
y : x a11 x
2 2 1
T
inner product y , y 1 2
y 1
y2 0
x1 , x 2 a11 x1 0.
Fall 2018 Prof.Jiang@ECE NYU 186
x1 , x 2 a11 x1 0 a11 : x1 , x 2 / x1 , x1
with D1 : x1 , x1 0.
1 2
Contradiction with x , x being
linearly independent. Thus,
x1 , x1 x1 , x 2
D2 : det 0.
x 2 , x1 x 2 , x 2
Fall 2018 Prof.Jiang@ECE NYU 190
So, we have obtained three mutually orthogonal
vectors:
y1 : x1
y : x a11 x
2 2 1
y : x a21 x a22 x
3 3 1 2
k
for any set of linearly independent vectors x i
.
i 1
i 1
j
u
j 1
x j
k
i , j 1
x , x ui u j
i j
a u
j 1
ij j 0, i 1, 2, , N
A 2
a21 a22 a
which is symmetric, i.e., a12 a21.
x12
Ax 1 x a , x 1 x11 ,
1 1 1 1
a , x 1 x12
2 1
1 y12 2
a , y 0 b22
2
Then, b12 0 using symmetry;
T
T
O AO2
2 O2T AO2 .
and b22 2 because the eigenvalues are
unchanged under O.
Fall 2018 Prof.Jiang@ECE NYU 203
Exercise 1
Try to reduce the real symmetric matrix
1 k
A
k 1
to a diagonal form.
See the text (Horn & Johnson, 2nd edition, 2013; page 226)
Fall 2018 Prof.Jiang@ECE NYU 205
Homework #4
1. Does the singular matrix
1 1
A
1 1
have two independent eigenvectors?