Scribd
Upload
18 views

Module 11 Assignment Solutions

Uploaded by

wes.s.devlin
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
18 views

Module 11 Assignment Solutions

Uploaded by

wes.s.devlin
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

Module 11 Assignment Solutions 1

Matrix Theory

***Please do not share or re-distribute these solutions!***

1. To find the eigenvalues of A, we need to compute

4−λ 4 1
3
χA (t) = (−1) det(A − λI ) = (−1) −1 −λ 0 .
0 0 2−λ

Applying co-factor expansion along the bottom row, we get

4−λ 4
χA (t) = −(2−λ) = −(2−λ) [(4 − λ)(−λ) − 4(−1)] = (λ−2)(λ2 −4λ+4) = (λ−2)3 .
−1 −λ

Thus the only eigenvalue of A is 2.


Next, we need to find an eigenvector associated with eigenvalue 2. To do so, we need to
find the null space of A − 2I . We begin by finding the row reduced echelon form of A − 2I :
         
2 4 1 1 2 12 1 2 21 1 2 12 1 2 0
 E1/2 [1]   E1 [2,1]   E2 [2]   E−1/2 [1,2] 
−1 −2 0 −−−→ −1 −2 0  −−−→ 0 0 12  −−→ 0 0 1  −−−−−→ 0 0 1 .
 
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Applying the rest of Gauss-Jordan elimination, we delete the last row, add a row of zeros
to the second row, and change the missing pivot to −1 to get
 
1 2 0
0 −1 0 .
 
0 0 1

This gives us the eigenvector u = (2, −1, 0)T associated wtih eigenvalue −1.
Next, we find a basis for C3×1 containing u by creating the matrix [u | I ] and reducing to
row reduced echelon form:
     
1 1
2 1 0 0 1 2
0 0 1 2
0 0
 E1/2 [1]   E1 [2,1]  1
−1 0 1 0 −−−→ −1 0 1 0 −−−→ 0 1 0
 
2
0 0 0 1 0 0 0 1 0 0 0 1
   
1 21 0 0 1 0 −1 0
E2 [2]   E−1/2 [1,2] 
−−→ 0 1 2 0 −−−−−→ 0 1 2 0 .

0 0 0 1 0 0 0 1
2

Since the first, second, and fourth columns are pivot columns, the corresponding
columns of [u | I ] form a basis for C3×1 : {u1 , u2 , u3 } = {(2, −1, 0)T , (1, 0, 0)T , (0, 0, 1)T }.
We can now use the Gram-Schmidt algorithm to generate an orthogonal basis.

v1 = (2, −1, 0)T


         
1 2 1 2 1/5
  hu2 , v1 i     2    
v2 = 0 − −1 = 0 − −1 = 2/5
hv1 , v1 i 5
0 0 0 0 0
       
0 2 1 0
  hu3 , v1 i   hu3 , v2 i    
v3 = 0 − −1 − 0 = 0
hv1 , v1 i hv2 , v2 i
1 0 0 1
Normalizing, we create the unitary matrix
 
√2 √1 0
 √51 √2
5
P= − 5 0 .

5
0 0 1
Then  
2 5 √2
5
∗ √1 
P AP = 0 2 =U
 
5
0 0 2
which is already upper triangular, so we are done! Our Schur triangularization is
 2   
√ √1 0 2 5 √2
 5 5 5
P = − √15 √25 0 and U = 0 2 √15  .
  

0 0 1 0 0 2

2. (a) We know by that for any n × n matrices A1 , A2 , tr(A1 A2 ) = tr(A2 A1 ) (see page 132 in
the textbook). Letting A1 = AB and A2 = C , observe that applying this property of
trace yields
tr(ABC ) = tr((AB)C ) = tr(C (AB)) = tr(C AB)
as desired.
(b) Let A ∈ Cn×n have Schur Triangularization U = P ∗ AP where P ∈ Cn×n is unitary
matrix and U ∈ Cn×n is upper triangular. By part (a),

tr(U) = tr(P ∗ AP) = tr(PP ∗ A) = tr(I A) = tr(A).

Since the elements on the main diagonal of U are the eigenvalues λ1 , . . . , λn of A,


we can conclude
n
X
tr(A) = tr(U) = λj .
j=1
3

Note. This argument can also be made using the Jordan Normal Form.
(c) Let A ∈ Cn×n have Schur Triangularization U = P ∗ AP where P ∈ Cn×n is unitary
matrix and U ∈ Cn×n is upper triangular. Because det(AB) = det(A)det(B) by
Theorem 3.30 (6),

det(U) = det(P ∗ AP) = det(P ∗ )det(A)det(P).

Since P is unitary, P ∗ = P −1 and by Theorem 3.30 (7), det(P ∗ ) = (det(P))−1 . That is,
det(P ∗ )det(P) = 1. Thus, det(U) = det(A).
Additionally, since the elements on the main diagonal of U are the eigenvalues
λ1 , . . . , λn of A and the determinant of a triangular matrix is the product of the
diagonal entries by Theorem 3.30 (5), we can conclude
n
Y
det(A) = det(U) = λj .
j=1

Note. This argument can also be made using the Jordan Normal Form.

3. Since null(T − 2I ) = 1 and null(T − 3I ) = 2, 2 has geometric multiplcity 1 and 3 has


geometric multiplicity 2. This tells us there is one Jordan block associated with
eigenvalue 2 and two Jordan blocks associated with eigenvalue 3.
Since χT (t) = (t − 2)2 (t − 3)3 , we know that T has eigenvalues 2 and 3, where the
algebraic multiplicity of 2 is 2 and the algebraic multiplicity of 3 is 3. This tells us the sum
of the sizes of the Jordan blocks associated with eigenvalue 2 is 2 (i.e., the Jordan block
associated with eigenvalue 2 is J(2, 2)) and the sum of the sizes of the Jordan blocks
associated with eigenvalue 3 is 3. The only possible way to write 3 as a sum of two
positive integers is 1 + 2, so the two Jordan blocks associated with eigenvalue 3 are
J(3, 1) and J(3, 2). Overall, this tells us the Jordan normal form of T (up to permutation of
the Jordan blocks) is
 
2 1 0 0 0
0 2 0 0 0
 
diag{J(2, 2), J(3, 1), J(3, 2)} = 0 0 3 0 0 .
 
 
0 0 0 3 1
0 0 0 0 3

You might also like