L2 - UE MAT334

Exercise sheet n◦ 7

Eigenvalues and eigenvectors

1. Compute the eigenvalues and the eigenvectors of the following matrices.

       
1 1 −1 −1 1 1 4 1 2 1 1 −2
 2 3 −4 
A= B= 1 −1 1  C= −1 1 −1 D =  0 2 −2  .
   
   
4 1 −4 1 1 −1 −2 −1 0 1 0 0

Which of the previous matrices is diagonalizable ?

2. Let  
3 −1 1
A= 0 2 0 

.
1 −1 3
(i) Show that A is diagonalizable and find a matrix P diagonalizing A. Compute An
for all n ≥ 1.

(ii) Consider the sequences (un ), (vn ) and (wn ) defined by the initial values u0 = v0 =
1, w0 = 2 and the following recursive relations :

un+1 = 3un − vn + wn vn+1 = 2vn wn+1 = un − vn + 3wn .

Compute un , vn and wn .
3. Let a, b ∈ IR and  
a b b
A =  b a b .
 

b b a
Compute An for all n ≥ 1.

4. Let A ∈ Mn (C) be a nilpotent matrix, i.e. there exists p ∈ N∗ such that Ap = 0. Show
that the only eigenvalue of A is 0. Does the converse hold ?

5. Let A ∈ Mn (C). Show that the determinant of A is equal to the product of its ei-
genvalues (counted with multiplicity) and that the trace of A is equal to the sum of its
eigenvalues (also counted with multiplicity).

6. Let A, B ∈ Mn (C). Show that AB and BA have the same set of eigenvalues, each of
them with the same multiplicity.

1
7. The Italian mathematician Leonardo Fibonacci (c. 1175 – c. 1250) was the first to
study the sequence of integers given by 1, 1, 2, 3, 5, 8,. . . and that has his name. It is
recursively defined by
f0 = 1, f1 = 1, fk+1 = fk + fk−1 , k = 1, 2, ...
1. Let x(k) = (fk+1 ; fk ). Write this relations in matricial form
x(k+1) = Ax(k) , k = 0, 1, ... , x(0) = (1; 1),
, where A is a matrix to be determiend .
2. Compute the eigenvalues and the eigenvectors of A. Is the matrix LA diagonali-
zable ?
3. Find an explicit formula for the k-th element of the Fibonacci sequence.
8. Compute the eigenvalues of the matrix
1 0 4 0
 
 0 1 0 4 
A= .
 
 1 0 1 0 
0 1 0 1
Consider now the system of differential equations given by
ẋ(t) = x(t) + 4z(t), ẏ(t) = y(t) + 4w(t),
ż(t) = x(t) + z(t), ẇ(t) = y(t) + w(t).
Compute the general solution of the system, and then the particular solution satisfying
the conditions x(0) = y(0) = z(0) = 0, w(0) = 2.

9. Let a ∈ IR. Consider the matrix

1 1 −1 0
 
 0 1 0 a 
A= .

0 −1 2 0 


1 0 1 2
(i) Compute the characteristic polynomial of A and the corresponding eigenvalues.

(ii) Compute the eigenspaces of A associated to each eigenvalue. Determine all the
values of the parameter a such that the matrix A is diagonalizable. And triangu-
larizable ?

(iii) Assume that a = 0 and consider the system of differential equations given by
ẋ(t) = x(t) + y(t) − z(t), ẏ(t) = y(t),
ż(t) = −y(t) + 2z(t), ẇ(t) = x(t) + z(t) + 2w(t).
Compute the general solution of the system, and then the particular solution sa-
tisfying the conditions x(0) = y(0) = w(0) = 0, z(0) = 1.

2
 tat.

Exercice 11. Deux masses égales sont suspendues entre trois ressorts identiques de coefficient de rigidité k sur une table plane
10. Consider
et lisse, comme the system
dans le diagramme formed by two masses and three springs with the same force
ci-dessous.
constant k > 0 as indicated in the following diagram :

Assume that the two are longitudinally separated from the equilibrium position (using an
external force) at time t = 0 and the systems stars moving freely. Let x1 (t) (resp. x2 (t))
the (signed) longitudinal distance of the first (resp. second) mass from its equilibrium
Les deux position
masses sont
at t.mises en mouvement
We assume entwo
that the temps t = 0. obey
springs Soit xHooke’s
1 (t ) (resp.law,
x2 (t ))
sol’écart de la première (resp. de la seconde)
masse de sa position d’équilibre en temps t . On admettra que les lois de Newton nous donnent que
mx001 = −2kx1 + kx2 ,
00 − 2kx .
mx002 =mxkx
1 = 2 kx2
1 −2kx1 +

1. Rewrite these equations in matrixmx

form
00 :
= kx1 − 2kx2 .
2
!00 !
1. Reécrire cette équation dans la forme x1 x1
µ= ¶M
00 µ ¶
x2 x1 x2x1
=M
x2 x2
where M s a matrix to be determined.
pour une certaine matrix M.
2. Diagonalize M and compute the general solution of this equation.
2. Diagonaliser M et donner la solution générale de cette équation.
3. What are normal modes of oscillation of the system ?
3. Interprêter physiquement les deux modes “fondamentales”, cad, les solutions particulières correspondantes à chaque va-
We consider now the analogous situation with 3 masses and 4 springs.
leur propre.
1. Write down the equations of motion of this system.
2. Rewrite them in matrix form :
 00  
x1 x1
 x2  = M  x2 
   

x3 x3

where M s a matrix to be determined.

3. Diagonalize M and compute the general solution of this equation.
4. What is the normal mode of oscillation of the system corresponding to the eigen-
value 2k/m.

3
Nous considérons maintenant la même situation avec 3 masses suspendues entre 4 ressorts, comme ci-dessous.

Eigenvalues of symmetric matrices and singular value decomposition

Nous admettrons les équations de mouvement,
1 mx 00 = −2kx + kx
1 2
11. Show that the following matrices are orthogonal and compute their eigenvalues.
  mx200 = kx1 − 2kx2 + kx3  
0 1 0 " # 0 1 0
100 1 1
(a)  −1 0 0  (b)mx
 
√3 = kx2 − 2kx3 . (c) 
 1 0 0 .

2 −1 1
0 −1
1. Re-écrire ces équations dans la0 forme 0 0 1
 00  
1 x 1 x
12. Show that the following matrices are symmetric and
 x 2  = M  x2 
diagonalize them by means of
an orhtogonal matrix.
x3 x3
   
pour une certaine matrix−2
"
M. 2
# " # 1 2 0 1 1 1
2 36
(a) (b) (c)  2 1 0  (d)  1 1 1 
   
2. Diagonaliser M et donner2la solution
1 générale
36 de23cette équation.
0 0 1 1 1 1
3. Interprêter physiquement la mode “fondamentale” correspondante à la valeur propre 2k/m.
13.lesCompute
4. Justifier équationsthe singular value
de mouvement decomposition
données ci-dessus. of the following matrices.
   
−2 0 1 0 0

" # " #
3 0 0 0 −2 0
(a) (b)  0 1  (c) (d)  0 0 0 
   
0 −1 0 2 0 0
0 −1 0 −1 2
Exercice 12. On cherche à déterminer les matrices X qui commutent avec la matrice
14. Let A = U ΣV t be the singular valuedecomposition  of a matrix A ∈ IR
m×n
of rank
2 0 4
p < min(m, n). We denote by σ1 ≥ σ2 ≥ ... ≥ σp the nonzero singular values of A, and
=  left
by Ū = [u1 , ..., up ] and V̄ = [v1 , ..., vp ]Athe 3 and 12  . associated orthogonal matrices,
−4 right
respectively. We have thus 1 −2 5
..., σp ) V̄ t vectoriel de M (R).
A = Ū diag(σ1 , sous-espace
- Montrer que les matrices X qui commutent avec A forment un n
−1
- Trouver une matrice diagonale
(minimal D et une matrice inversible P telles que P AP = D.
decomposition).
- Déterminer lesLet
matrices
A+ beY the
qui commutent avecnD×etmengiven
matrix of size déduire
by celles qui commutent avec A.

A+ = V̄ diag(σ1−1 , ..., σp−1 ) Ū t .


It is called the pseudo-inverse matrix of A.
Exercice 13. Soit A ∈ Mn (C) une matrice nilpotente, c’est-à-dire qu’il existe p ∈ N∗ tel que Ap = 0. Montrer que 0 est l’unique
valeur propre de A. La réciproque est-elle vraie ? 4


1. What is the size of A+ ? Express A+ A and AA+ in terms of the singular value
decomposition of A. Verify that A A+ A = A and A+ A At = At AA+ = At . Explain
their meaning.
2. Let A be a matrix of size m × n and rank n, and let b be a vector column of size m.
Assume that the system Ax = b has a solution x. Show that the solution x ∈ IRn
satisfies x = A+ b.