Eigenvectors and eigenvalues Page 1

Eigenvectors and eigenvalues

There is an important aspect of matrix theory that arises from the fact that matrices
can be used to represent transformations. If a vector v is transformed by a matrix A,
then the resulting vector is Av. If the direction of the vector is unchanged once it has
been transformed we can write:

Av   v

where  is a constant. The vector v is called an ‘eigenvector’ of the matrix, and the
corresponding value of  is called an ‘eigenvalue’.

To find the eigenvectors and eigenvalues, you have to work with the equation Av   v :

Av   v  (A  I)v  0

where 0 is the zero matrix. Notice that we had to insert the identity matrix I into the
equation since we cannot subtract a scalar from a matrix.

This equation would be true if either (A  I)  0 , or v  0 , but these are not going to be
very helpful since this is just the trivial solution and is not the one we are interested in.

To solve a general matrix equation Bx  C , we would normally find the inverse of B and
calculate x  B 1C . Using this technique to try to solve the equation (A  I)v  0 , then
if (A  I) has an inverse we get v  (A  I)1 0 , ie v  0 . Since we are looking for a non-
trivial solution, we must prevent this method from working. The only thing that would
stop us getting v  0 would be if A  I does not have an inverse ie it is singular.
Remembering that singular matrices have a determinant of zero this gives us a way to
find the eigenvalues and eigenvectors.

In summary, to find the eigenvalues of a matrix A we must solve the equation

det(A  I)  0 . The equation obtained from det(A  I)  0 is called the ‘characteristic
equation of the matrix’.

Once you have the eigenvalues, you can return to the equation (A  I)v  0 , to find the
eigenvectors.

This method will work for a general n  n matrix, but due to the complication of
calculating the determinant we will only look at 2  2 and 3  3 matrices here.

The Actuarial Education Company © IFE: 2019 Examinations

Page 2 Eigenvectors and eigenvalues

Example

 2 1 
Find the eigenvectors and corresponding eigenvalues of the matrix B   .
2 5

Solution

First we use the equation det(B  I)  0 to find the eigenvalues:

 2   1 
det  0
 2 5 

(2   )(5   )  2  0

 2  7  12  0
(  3)(  4)  0

so   3 or   4 .

x
We now use the equation (B  I)v  0 to find the eigenvectors, letting v    . There is
y
a separate eigenvector corresponding to each eigenvalue.

2   1  x   0   1 1  x   0   x  y   0 
If   3 , then       , ie       , ie      . Any
 2 5    y   0   2 2  y   0   2 x  2y   0 
value of x and y such that x  y  0 ie y   x will make both rows work out.

1
So the eigenvector corresponding to   3 is k   , where k is a constant.
 1 

 2 1  x   0   1
If   4 , then       , so the eigenvector corresponding to   4 is k   .
2 1  y   0   2 

Notice that we include k in the eigenvector, since any value of the constant k will give
an eigenvector.

© IFE: 2019 Examinations The Actuarial Education Company

Eigenvectors and eigenvalues Page 3

Question 1.1

1 2 
Find the eigenvectors and corresponding eigenvalues of the matrix B   .
 3 4 

Question 1.2

 2 2 3 
Show that   1 is an eigenvalue of the matrix  1 1 1  , and hence find all the
 
 1 3 1 
eigenvalues and eigenvectors.

You are given that  3  2 2  5  6  (  1)( 2    6) .

The Actuarial Education Company © IFE: 2019 Examinations

Page 4 Eigenvectors and eigenvalues

Solutions

Solution 1.1

1   2 
For eigenvalues, det    0 , so we have the characteristic equation:
 3 4   

(1   )(4   )  6  0   2  3  10  0

This gives   2,   5 .

1   2  x   0   1 2  x   0 
If   2 , then              .
 3 4    y   0   3 6  y   0 

2
So the eigenvector corresponding to   2 is k   .
1

1   2  x   0   6 2  x   0 
If   5 , then              .
 3 4    y   0   3 1  y   0 

1
So the eigenvector corresponding to   5 is k   .
 3 

© IFE: 2019 Examinations The Actuarial Education Company

Eigenvectors and eigenvalues Page 5

Solution 1.2

For the eigenvalues we must solve the equation:

 2   2 3 
 
det  1 1 1 0
 1 3 1   


So we have the equation:

(2   )(1   )(1   )  3  2(1   )  1  33  (1   )  0

and this equation simplifies to  3  2 2  5  6  0 .

Factorising the given cubic equation completely gives:

(  1)(  3)(  2)  0

The eigenvalues are 1, 3, and –2.

 2   2 3  x   0   1 2 3  x   0 
When   1 ,  1 1           
1  y    0    1 0 1  y    0  , which gives the
 1 1           
 3  z   0   1 3 2  z   0 
 1 
eigenvector of k  1  .
1
 

 2   2 3  x   0   1 2 3  x   0 
When   3 ,  1 1           
1  y    0    1 2 1  y    0  , which gives the
 1 1           
 3  z   0   1 3 4  z   0 
1
eigenvector of k  1  .
1
 

The Actuarial Education Company © IFE: 2019 Examinations

Page 6 Eigenvectors and eigenvalues

 2   2 3  x   0   4 2 3  x   0 
When   2 ,  1 1           
1  y    0    1 3 1  y    0  , which gives the
 1 3 1           
  z   0   1 3 1  z   0 
 11 
eigenvector of k  1  .
 14 
 

© IFE: 2019 Examinations The Actuarial Education Company