Skip to main content
Logo image

Section 21.5 Examples

Here we will compute eigenvalues and a basis for each corresponding eigenspace for the matrices in Discovery 21.3.

Example 21.5.1. A 2×2 example.

First, we form the matrix
λIA=[λ784λ+5].
Then we compute its determinant, to obtain the characteristic polynomial of A:
cA(λ)=det(λIA)=(λ7)(λ+5)+32=λ22λ3=(λ+1)(λ3).
The eigenvalues are the roots of the characteristic polynomial, so we have two eigenvalues λ1=1 and λ2=3.
The eigenspace Eλ1(A) is the same as the null space of the matrix λ1IA, so we determine a basis for the eigenspace by row reducing:
(1)IA=[8844]reducerow[1100].
This system requires one parameter to solve, as x2 is free. Setting x2=t, the general solution in parametric form is
x=[x1x2]=[tt]=t[11].
Associated to the single parameter we get a single basis vector, so that
dim(Eλ1(A))=1.
In particular, we have
Eλ1(A)=Span{[11]}.
Now move on to the next eigenvalue. Again, we determine a basis for Eλ2(A) by row reducing λ2IA:
3IA=[4848]reducerow[1200].
Again, x2 is free. One parameter means one basis vector, so again
dim(Eλ2(A))=1.
The first row of the reduced matrix says x1=2x2, so we have
Eλ2(A)=Span{[21]}.

Example 21.5.2. A 3×3 example.

Start with
λIA=[λ2440λ+6806λ8],
and compute the characteristic polynomial,
cA(λ)=det(λIA)=(λ2)[(λ+6)(λ8)+48]=(λ2)(λ22λ)=λ(λ2)2.
The eigenvalues are λ1=0 and λ2=2.
The eigenspace Eλ1(A) is the null space of 0IA=A, so row reduce:
0IA=[244068068]reducerow[102/3014/3000].
Notice that the null space of 0IA=A is the same as the null space of A, since our first step in row reducing A could be to multiply each row by 1. Since this homogeneous system has nontrivial solutions, A must be singular.
The homogeneous system (λ1IA)x=0 requires one parameter, so
dim(Eλ1(A))=1.
The variable x3 is free, and the nonzero rows of the reduced matrix tell us x1=(2/3)x3 and x2=(4/3)x3. Setting x3=t, our general solution in parametric form is
x=[x1x2x3]=[(2/3)t(4/3)tt]=t[2/34/31].
However, to avoid fractions in our basis vector, we may wish to pull out an additional scalar:
x=t3[243],
giving us
Eλ1(A)=Span{[243]}.
Now row reduce λ2IA:
2IA=[044088066]reducerow[011000000].
This time we have two free variables, so dim(Eλ2(A))=2. Setting x1=s and x3=t, the general solution in parametric form is
x=[x1x2x3]=[stt]=s[100]+t[011],
giving us
Eλ2(A)=Span{[100],[011]}.

Example 21.5.3. A diagonal example.

This time our matrix is diagonal, so its eigenvalues are precisely the diagonal entries, λ1=1, λ2=2, λ3=3.
As usual, analyze each eigenvalue in turn.
For λ=1:
1IA=[000010002]reducerow[010001000]Eλ1(A)=Span{[100]}.
For λ=2:
2IA=[100000001]reducerow[100001000]Eλ2(A)=Span{[010]}.
For λ=3:
3IA=[200010000]reducerow[100010000]Eλ3(A)=Span{[001]}.
The fact that the eigenvectors of our diagonal matrix are standard basis vectors shouldn’t be too surprising, since a matrix times a standard basis vector is equal to the corresponding column of the matrix, and the columns of a diagonal matrix are scalar multiples of the standard basis vectors.

Example 21.5.4. An upper triangular example.

Our final example matrix is upper triangular, so again its eigenvalues are precisely the diagonal entries, λ1=2 and λ2=1.
Note that we don’t count the repeated diagonal entry 2 as two separate eigenvalues — that eigenvalue is just repeated as a root of the characteristic polynomial. (But this repetition will become important in the next chapter.)
Once again we determine eigenspaces by row reducing, one at a time.
For λ1=2:
2IA=[010000003]reducerow[010001000]Eλ1(A)=Span{[100]}.
For λ2=1:
(1)IA=[310030000]reducerow[100010000]Eλ2(A)=Span{[001]}.

Example 21.5.5. Using row operations to help.

Don’t forget that we can use row operations to help compute determinants!
Let’s do a 4×4 example to demonstrate. Consider
A=[542746211220228141135].
To obtain the characteristic polynomial, we want to compute the determinant of
λIA=[λ5427462λ+1122022λ+814113λ5].
Let’s row reduce a bit first:
[λ5427462λ+1122022λ+814113λ5]xR1R4x[1135λ2λ+1122022λ+814λ542746]xR2+2R1R3+2R1R4(λ5)R1[1135λ0λ162(λ+5)00λ+22(λ+2)0λ13(λ+4)λ210λ21].
In our first step above, we performed two operations: swapping rows and multiplying a row by 1. Both of these operations change the determinant by a factor of 1, so the two effects cancel out. Our other operations in the second step above do not affect the determinant, so the determinant of this third matrix above will be equal to the characteristic polynomial of A.
Now, we cannot divide a row by zero. So we should not divide either the second or fourth rows by λ1 in an attempt to obtain the next leading one, because we would inadvertently be dividing by zero in the case λ=1. However, we can still simplify one step further, even without a leading one:
[1135λ0λ162(λ+5)00λ+22(λ+2)0λ13(λ+4)λ210λ21]xxxR4R2(✶)[1135λ0λ162(λ+5)00λ+22(λ+2)003(λ+2)λ28λ11].
This last matrix is not quite upper triangular, but it’s close enough that we can proceed by cofactors from here.
cA(λ)=|1135λ0λ162(λ+5)00λ+22(λ+2)003(λ+2)λ28λ11|=1|λ162(λ+5)0λ+22(λ+2)03(λ+2)λ28λ11|=(λ1)|λ+22(λ+2)3(λ+2)λ28λ11|=(λ1)((λ+2)(λ28λ11)+6(λ+2)2)=(λ1)(λ+2)((λ28λ11)+6(λ+2))=(λ1)(λ+2)(λ22λ+1)=(λ1)(λ+2)(λ1)2=(λ1)3(λ+2).
We now see that the eigenvalues are λ1=1 and λ2=2.
To determine bases for eigenspaces, we usually reduce the matrix λIA with the various eigenvalues substituted in for λ. But we have already partially reduced λIA with λ left variable to help us determine the eigenvalues. So we can begin from (✶) for both eigenvalues.
For λ1=1:
[113400612003600918]reducerow[1102001200000000]Eλ1(A)=Span{[2021],[1100]}.
For λ2=2, again starting from (✶):
[1137036600000009]reducerow[1050012000010000]Eλ2(A)=Span{[5210]}.