MINIMAL POLYNOMIAL

Definition: Let f (a)  an x n  an1 x n1  an2 x n2    a2 x 2  a1 x  a0 be a polynomial of

degree n. Then an is called the leading coefficient and f (x) is monic if an  1.

Example: x 3  x 2  1, x 2  1 and x 4  2 x 3  3x  1 are all monic polynomials.

Definition: Let A be a square matrix. Then the minimal polynomial of A is a monic

polynomial M (x) of lowest degree such that M ( A)  0.

Theorem: The minimal polynomial of a matrix divides the characteristic polynomial.

Theorem: The minimal polynomial and the characteristic polynomial have the same
irreducible factors (factors that cannot be factorized further)
eg:

If D( x)  ( x  1) 2 ( x  2)( x  3)

M ( x)  ( x  1)( x  2)( x  3)
or
M ( x)  ( x  1) 2 ( x  2)( x  3) are the min imal polynomials.

Example 1: Find the characteristic polynomial, minimal polynomial and eigenvalues

of the matrix:

 3  5 5
 
A   5  7 5
 5  5 3
 
Solution:

x5 5 5
x7 5 5 5 5 x7
xI  A   5 x7  5  ( x  3) 5 5
5 x3 5 x3 5 5
5 5 x3

 ( x  3)( x  7)( x  3)  25  5 5( x  3)  25  5 25  5( x  7)

 ( x  3)( x  2) 2
Therefore, the characteristic polynomial of A is D( x)  ( x  3)( x  2) 2 .
The eigenvalues of A are   2,2 and   3.

The possibilities of minimal polynomial are:

M 1 ( x)  ( x  3)( x  2)

M 2 ( x)  ( x  3)( x  2) 2

 0  5 5  5  5 5   0 0 0 
    
M 1 ( A)  ( A  3I )( A  2 I )   5  10 5  5  5 5    0 0 0   0
 5  5 0  5  5 5   0 0 0 
    
M 2 ( A)  0 by Cayley-Hamilton theorem.  M 1 ( x )  ( x  3)( x  2) is the minimal
polynomial.
Example 2: Find the characteristic polynomial, minimal polynomial and the
eigenvalues of

2 1 0 0
 
0 2 0 0
A
0 0 2 0
 
0 0 0 5 

Solution:

x2 1 0 0
0 x2 0 0 x2 1 x  2 0
xI  A  
0 0 x2 0 0 x2 0 x 5
0 0 0 x 5

 ( x  2) 2 ( x  2)( x  5)  ( x  2) 3 ( x  5)
Therefore the characteristic polynomial of A is f ( x)  ( x  2) 3 ( x  5) .
The eigenvalues of A are   2,2,2 and   5.

Possibilities of minimal polynomial are

M 1 ( x)  ( x  2)( x  5)
M 2 ( x)  ( x  2) 2 ( x  5)
M 3 ( x)  ( x  2) 3 ( x  5)

0 1 0 0   3 1 0 0 0  3 0 0
    
0 0 0 0  0  3 0 0 0 0 0 0
M 1 ( A)  ( A  2 I )( A  5I )    0
0 0 0 0  0 0 3 0 0 0 0 0
    
0  0   0 0 0 0 
 0 0 3  0 0 0

Hence M 1 ( x ) is not a minimal polynomial of A.

2
0 1 0 0 3 1 0 0  0 0 0 0   3 1 0 0
      
0 0 0 0  0 3 0 0  0 0 0 0  0  3 0 0
M 2 ( A)  ( A  2 I ) ( A  5 I )  
2

0 0 0 0  0 0 3 0  0 0 0 0  0 0 3 0
      
0 0 0 3   0 0   0  0 
  0 0 0 0 9  0 0 0

0 0 0 0
 
0 0 0 0
 0
0 0 0 0
 
0 0 0 0 


M 3 ( A)  0 by Cayley-Hamilton theorem. Hence M 2 ( x)  ( x  2) 2 ( x  5) is the minimal

polynomial of A.
Exercise:

1. For each of the following matrix find the polynomial having the matrix as a root,
 1 4  3
2 5   2  3  
i. A    ii. B    iii. C   0 3 1 
 1  3  7  4  0 2  1
 
2. Find the characteristic polynomial, minimal polynomial and eigenvalues of each
matrix below;
2 1 00
1 1 0   1 2 2
  0 2 0 0  
i. A  0 2 0 ii. B   iii. C   1 2  1
0 0 1 0 0 1 1  1 1 4 
     
0 0  2 4 


MINIMAL POLYNOMIAL THEOREMS

Theorem: The minimal polynomial M (x) of a matrix A divides every polynomial which
has A as a zero. In particular, M (x) divides the characteristic polynomial of A.

Proof: Suppose f (x) is a polynomial for which f ( A)  0. By division algorithm there

exists polynomials q(x) and r (x) for which f ( x)  M ( x)q( x)  r ( x) and r ( x)  0 or
deg r ( x)  deg M ( x). Substituting x  A in this equation and using the fact that f ( A)  0
and M ( A)  0, we obtain r ( A)  0. If r ( A)  0, then r (x) is a polynomial of degree less
than M (x) which has A as a zero; this contradicts the definition of a minimal
polynomial. Thus r ( x)  0 and so f ( x)  M ( x)q( x) i.e M (x) divides f (x).

Theorem: Let M (x) be the minimal polynomial of an n-square matrix A. Then the
characteristic polynomial of A divides M (x).

Theorem: The characteristic polynomial D(x) and the minimal polynomial M (x) of a
matrix A have the same irreducible polynomials.

Proof: Suppose f (x) is an irreducible polynomial. If f (x) divides M (x) then since
M (x) divides D(x), f (x) divides D(x) . On the other hand if f (x) divides D(x) then
n
by the preceding theorem, f (x) divides ( M ( x)) . But f (x) is irreducible, hence f (x)
also divides M (x). Thus M (x) and D(x) have the same irreducible factors.
TRACE

The trace of a square matrix A  aij  written tr (A) is defined to be the sum of its

diagonal elements i.e tr ( A)  a11  a22  a33    ann

Theorem: Similar matrices have the same trace.

Thus the trace of a linear operator T is the trace of any one of its matrix
representations i.e tr (T )  tr T e .

1 2
Example: Let V be the space of 2 2 matrices over ℝ , and let M   . Let T be the
3 4
linear operator on V defined by T ( A)  MA. Find the trace of T.

Solution: We must find a matrix of representation of T. Choose the usual basis of V

 1 0 0 1  0 0  0 0 
 E1   , E2   , E3   , E4    , then
  0 0  0 0 1 0  0 1 

 1 2  1 0   1 0 
T ( E1 )         1E1  0 E2  3E3  0 E4
 3 4  0 0   3 0 

 1 2  0 1   0 1 
T ( E2 )         0 E1  1E2  0 E3  3E4
 3 4  0 0   0 3 

 1 2  0 0   2 0 
T ( E3 )         2 E1  0 E2  4 E3  0 E4
 3 4  1 0   4 0 

 1 2  0 0   0 2 
T ( E4 )         0 E1  2 E2  0 E3  4 E4
 3 4  0 4   0 4 
Hence

1 0 2 0
 
0 1 0 2
T E  and tr (T )  1  1  4  4  10
3 0 4 0
 
0 3 0 4 
