EE5413: Linear Dynamical Systems

Assignment 2
August 18, 2018

1. Consider a state-space realization for a system as follows:

(A, B, C, D) : ẋ = Ax + Bu, y = Cx + Du

Find an expression for the transfer fuction of the above system in terms of A, B, C&D. Show that the
transfer function obtained for the realization (Ā, B̄, C̄, D̄), where Ā = T −1 AT , B̄ = T −1 B, C̄ = CT and
D̄ = D is the same as the transfer function obtained for the above realisation.
2. (a) For A ∈ Rn×n , let A = T −1 JT , where J is the Jordan normal form of matrix A. Show that
Ak = T −1 J k T.
(b) Let
 
λ 1 0 ··· 0
0 λ 1 ··· 0
J = .
 
.. .. .. 
 .. . . ··· .
0 0 0 ··· λ

be of size n. Write an expression for J k .

2 2
A3 t3
(c) For A ∈ Rn×n , eAt = I + At + A2!t + 3! + · · · . Use this and the expression for J k found in (b) to
determine an expression for eJt .
(d) Consider a block diagonal matrix
 
B1
 B2 0 
B= .
 
..
.
0
 
Bm

Each block Bi is a square matrix of size ni . Show that

 k 
B1
 B2k 0 
Bk =  .
 
..
.
0
 
k
Bm

(e) Let A have the Jordan canonical form

 
J1
 J2 0 
J = ,
 
..
.
0
 
Jp

where Ji are the Jordan blocks of A. Show that

 Jt 
e 1
 eJ 2 t 0 
eAt = T −1   T.
 
..
.
0
 
eJ p t

1
3. (a) Find the eigenvalues and eigenvectors of the matrices
     
1 −3 3 3 −1 1 2 0 2
A1 = −2 0 2 , A2 = −1 3 1 and A3 = −1 3 1 .
1 −1 3 0 0 4 −1 1 5
and comment on the diagonalizability of the matrices.
(b) If a matrix is diagonalizable, find the transformation matrix and the diagonal matrix. If not, find
the Jordan canonical form.
(c) Using the results shown in Question 2., find the matrix exponential of each matrix.
Note :- The computation of eigenvalues and eigenvectors for the given three matrices, involves simple hand
calculations. However, you can verify your computations using the MATLAB commands :
(a) [V,D]=eig(A) to find the eigenvalues (diagonal entries of D) and eigenvectors (columns of V) for a
square matrix A.
(b) Comment on the returns of the eig function, when the matrix A is not diagonalizable.
(c) Use [V,J]=jordan(A) to find the Jordan canonical form J and the transformation matrix V for a
square matrix A.
4. Check if the following two representations are equivalent.
R1 :
   
−1 0 1
ẋ1 (t) = x (t) + u(t),
0 −2 1 −1
y(t) = [1 1] x1 (t) + u(t).

R2 :
   
0 1 0
ẋ2 (t) = x (t) + u(t),
−2 −3 2 1
y(t) = [1 0] x2 (t).

5. Check if the following statements are true or false. If false, give the correct statement.
(a) Two matrices are similar if they have the same eigenvalues.
(b) Similar matrices have the same eigenvectors.
6. Determine the characteristic polynomial and minimal polynomial of the following matrices. Note that the
matrices are in Jordan form.
   
1 1 0 0 1 1 0 0
0 1 1 0 0 1 0 0
A1 = 0 0 1 0 ,

0 0 1 0 ,
A2 =  

0 0 0 1 0 0 0 1
 
1 1 0 0
0 1 0 0
A3 = 0 0 1 1 ,
 A4 = I4 .
0 0 0 1

7. Check if the following matrices are cyclic. If yes, find a cyclic generator for each.
   
1.5 0.5 0 1.5 0.5 0
A1 = −0.5 0.5 0 , A2 =  0.5 1.5 0 ,
−1.5 −0.5 0 −0.5 −0.5 1
 
1.5 0.5 0
A3 = −0.5 0.5 0 .
−0.5 −0.5 1
Hint: Use the transformation matrix,
 
1 −1 0
T = 1 1 0 ,
1 0 1
to get the Jordan form J = T AT −1 .

2
8. Prove that two matrices A ∈ Rn×n and B ∈ Rn×n are similar only if they have the same characteristic
polynomial and minimal polynomial. Check if this condition is sufficient also for the case when n ≤ 3.
9. Let λ be a scalar such that (C − λI)n×n is singular.

(a) If a matrix B is similar to C, prove that (B − λI) is also singular.

(b) Prove that (B − λi I) is singular whenever Bn×n is similar to D = diag{λ1 , λ2 , . . . , λn }.