Eigenvalues and eigenvectors

Eigenvalues and eigenvectors

Spec 3.1 · Edexcel A-Level Further Maths: Revision notes

Studying this course? Keep your progress in one place.

Key facts

  • An eigenvector v\mathbf{v} of A\mathbf{A} satisfies Av=λv\mathbf{A}\mathbf{v}=\lambda\mathbf{v}; λ\lambda is the eigenvalue. The zero vector is never an eigenvector.
  • Eigenvalues solve the characteristic equation det⁡(A−λI)=0\det(\mathbf{A}-\lambda\mathbf{I})=0.
  • For 2×22\times2: λ2−(trace)λ+det⁡=0\lambda^2-(\text{trace})\lambda+\det=0. Eigenvalues sum to the trace and multiply to the determinant.
  • For each λ\lambda, solve (A−λI)v=0(\mathbf{A}-\lambda\mathbf{I})\mathbf{v}=\mathbf{0} and check Av=λv\mathbf{A}\mathbf{v}=\lambda\mathbf{v}.
  • Anv=λnv\mathbf{A}^n\mathbf{v}=\lambda^n\mathbf{v}.

Eigenvalues and eigenvectors

An eigenvector keeps its direction under the matrix; the eigenvalue is the scale factor.

A non-zero vector v\mathbf{v} is an eigenvector of a square matrix A\mathbf{A} if Av=λv\mathbf{A}\mathbf{v}=\lambda\mathbf{v} for some scalar λ\lambda, the eigenvalue.

Geometrically, A\mathbf{A} maps the line through the origin in the direction of v\mathbf{v} onto itself, scaling it by λ\lambda (and reversing it if λ<0\lambda<0). This line is an invariant line.

Any non-zero multiple of an eigenvector is also an eigenvector.

vAv = 4vOVW
v = (1, 1) maps to (4, 4) = 4v: same direction, scaled by 4.
  • Eigenvector equationAv=λv\mathbf{A}\mathbf{v}=\lambda\mathbf{v}

Which of these is an eigenvector of (3122)\begin{pmatrix}3&1\\2&2\end{pmatrix}?

The characteristic equation

Eigenvalues are the roots of det⁡(A−λI)=0\det(\mathbf{A}-\lambda\mathbf{I})=0.

Av=λv\mathbf{A}\mathbf{v}=\lambda\mathbf{v} means (A−λI)v=0(\mathbf{A}-\lambda\mathbf{I})\mathbf{v}=\mathbf{0}, which has a non-zero solution only if det⁡(A−λI)=0\det(\mathbf{A}-\lambda\mathbf{I})=0. This is the characteristic equation.

For a 2×22\times2 matrix it is λ2−(trace A)λ+det⁡A=0\lambda^2-(\text{trace }\mathbf{A})\lambda+\det\mathbf{A}=0, where the trace is the sum of the leading-diagonal entries.

The eigenvalues sum to the trace and multiply to the determinant, a quick check. A matrix is singular exactly when 00 is an eigenvalue.

−1123456−2246xyλ = 1λ = 4λ² − 5λ + 4
The eigenvalues are where λ² − 5λ + 4 crosses zero.
  • 2×22\times2 characteristic equationλ2−(trace)λ+det⁡=0\lambda^2-(\text{trace})\lambda+\det=0

Worked example

Find the eigenvalues of A=(3122)\mathbf{A}=\begin{pmatrix}3&1\\2&2\end{pmatrix}.

A 2×22\times2 matrix has trace 77 and determinant 1010. What are its eigenvalues?

Finding eigenvectors

Solve (A−λI)v=0(\mathbf{A}-\lambda\mathbf{I})\mathbf{v}=\mathbf{0}: the two equations are multiples, so choose any convenient non-zero solution.

For each eigenvalue solve (A−λI)v=0(\mathbf{A}-\lambda\mathbf{I})\mathbf{v}=\mathbf{0}. The two equations are multiples of each other (that is why λ\lambda is an eigenvalue), so you get one equation in xx and yy and choose a convenient non-zero multiple.

For A=(3122)\mathbf{A}=\begin{pmatrix}3&1\\2&2\end{pmatrix}: λ=4\lambda=4 gives −x+y=0-x+y=0, so v=(11)\mathbf{v}=\begin{pmatrix}1\\1\end{pmatrix}; λ=1\lambda=1 gives 2x+y=02x+y=0, so v=(1−2)\mathbf{v}=\begin{pmatrix}1\\-2\end{pmatrix}.

A normalised vector has magnitude 11: divide by ∣v∣|\mathbf{v}|.

  1. 1

    Find the eigenvalue

    from the characteristic equation

  2. 2

    Form A − λI

    subtract λ from each diagonal entry

  3. 3

    Solve (A − λI)v = 0

    one equation in x and y

  4. 4

    Choose a multiple

    and check that Av = λv

Method.

Worked example

Find the eigenvalues and eigenvectors of P=(2112)\mathbf{P}=\begin{pmatrix}2&1\\1&2\end{pmatrix}.

Normalise the eigenvector (34)\begin{pmatrix}3\\4\end{pmatrix}.

Repeated eigenvalues

A repeated eigenvalue may give one eigenvector direction, or every direction for a multiple of the identity.

If the characteristic equation has a repeated eigenvalue, λ\lambda has multiplicity 22.

For C=(3−111)\mathbf{C}=\begin{pmatrix}3&-1\\1&1\end{pmatrix}: λ2−4λ+4=(λ−2)2=0\lambda^2-4\lambda+4=(\lambda-2)^2=0, so λ=2\lambda=2 twice. Solving (C−2I)v=0(\mathbf{C}-2\mathbf{I})\mathbf{v}=\mathbf{0} gives x=yx=y: a single direction (11)\begin{pmatrix}1\\1\end{pmatrix}.

Do not assume a repeated eigenvalue gives two independent eigenvectors.

the only eigenvector lineOv = (1, 1)
C = (3 −1; 1 1) has the repeated eigenvalue 2 and only one eigenvector direction, (1, 1): C(1, 1) = (2, 2).

Typical case

  • e.g. (2102)\begin{pmatrix}2&1\\0&2\end{pmatrix}
  • One eigenvector direction
  • Only multiples of one vector

Multiple of the identity

  • e.g. 3I3\mathbf{I}
  • Every non-zero vector is an eigenvector
  • Two independent eigenvectors

The matrix (2102)\begin{pmatrix}2&1\\0&2\end{pmatrix} has eigenvalue 22 twice. How many independent eigenvector directions does it have?

Complex eigenvalues

A real matrix with a negative discriminant has complex eigenvalues in conjugate pairs, with conjugate eigenvectors.

If the discriminant of the characteristic equation is negative, the eigenvalues are complex. For a real matrix they come as a conjugate pair, and the eigenvectors are complex conjugates too.

For B=(1−221)\mathbf{B}=\begin{pmatrix}1&-2\\2&1\end{pmatrix}: (1−λ)2+4=0(1-\lambda)^2+4=0, so λ=1±2i\lambda=1\pm2i.

For λ=1+2i\lambda=1+2i: −2ix−2y=0-2ix-2y=0, so v=(1−i)\mathbf{v}=\begin{pmatrix}1\\-i\end{pmatrix}. For λ=1−2i\lambda=1-2i the eigenvector is the conjugate, (1i)\begin{pmatrix}1\\i\end{pmatrix}.

Geometrically B\mathbf{B} is a rotation with an enlargement, so no real line is invariant.

√563.4°O(1, 0)(1, 2)
B = (1 −2; 2 1) sends (1, 0) to (1, 2): a rotation by 63.4° with enlargement by √5, so no real direction is kept.
  • Discriminant < 0λ=p±qi\lambda=p\pm qi
  • B=(1−221)\mathbf{B}=\begin{pmatrix}1&-2\\2&1\end{pmatrix}λ=1±2i\lambda=1\pm2i

What are the eigenvalues of the rotation (0−110)\begin{pmatrix}0&-1\\1&0\end{pmatrix} through 90∘90^\circ?

Powers and larger matrices

If Av=λv\mathbf{A}\mathbf{v}=\lambda\mathbf{v} then Anv=λnv\mathbf{A}^n\mathbf{v}=\lambda^n\mathbf{v}, so write a vector in terms of eigenvectors.

To find Anw\mathbf{A}^n\mathbf{w} for any vector w\mathbf{w}, write w\mathbf{w} as a combination of eigenvectors and scale each part by its λn\lambda^n.

Compute λn\lambda^n for each eigenvalue separately; do not raise the whole matrix to a power.

For a 3×33\times3 matrix the characteristic equation is a cubic, solved by factorising. Each eigenvector comes from solving (A−λI)v=0(\mathbf{A}-\lambda\mathbf{I})\mathbf{v}=\mathbf{0} with three equations.

0.511.522.533.544.55246810xyλ = 2λ = 0.5
A component along an eigenvector is scaled by λⁿ: it grows for λ = 2 and shrinks for λ = 0.5 (illustrative).
  • Av=λv\mathbf{A}\mathbf{v}=\lambda\mathbf{v}Anv=λnv\mathbf{A}^n\mathbf{v}=\lambda^n\mathbf{v}

Worked example

M=(5234)\mathbf{M}=\begin{pmatrix}5&2\\3&4\end{pmatrix} has eigenvalues 77 and 22 with eigenvectors (11)\begin{pmatrix}1\\1\end{pmatrix} and (2−3)\begin{pmatrix}2\\-3\end{pmatrix}. Find Mn(4−1)\mathbf{M}^n\begin{pmatrix}4\\-1\end{pmatrix}.

A\mathbf{A} has eigenvalue 33 with eigenvector v\mathbf{v}. What is A4v\mathbf{A}^4\mathbf{v}?

Try an exam question

Find the eigenvalues of A=(1232)\mathbf{A}=\begin{pmatrix}1&2\\3&2\end{pmatrix} and a corresponding eigenvector for each eigenvalue.

[5 marks]

That's the notes covered.

Carry on to the next subtopic.

Try a question

Which is the characteristic equation of A\mathbf{A}?

1 mark, marked against the mark scheme

Revision notes