Diagonalisation and the Cayley-Hamilton theorem
Studying this course? Keep your progress in one place.
Key facts
- With eigenvectors as the columns of : , the diagonal matrix of eigenvalues in the same order.
- Powers: .
- A symmetric matrix has real eigenvalues and perpendicular eigenvectors, so with orthogonal.
- A matrix is not diagonalisable without two independent eigenvectors.
- Cayley-Hamilton: , with the trace and the determinant.
Diagonalising a matrix
Put the eigenvectors in the columns of and the eigenvalues, in the same order, on the diagonal of .
If has eigenvalues with independent eigenvectors , form with the eigenvectors as columns. Then , a diagonal matrix.
The order of the eigenvalues in must match the order of the eigenvectors in . Swapping columns of swaps the entries of .
Check to avoid finding . A matrix is diagonalisable when this works.
- eigenvectors as columns
Worked example
Diagonalise .
- 1
Trace , determinant : , so or .
- 2
gives , so .
- 3
gives , so .
- 4
Put the eigenvectors in the columns of and the eigenvalues in the same order in .
has eigenvalue with eigenvector and eigenvalue with eigenvector . If , what is ?
Powers of a matrix
Since , the inner factors cancel and .
Since , each in a product cancels, leaving .
Powering a diagonal matrix just powers each entry: .
Always check returns the original matrix.
- λ = 2
- λ = 3
Worked example
Find for .
- 1
Eigenvalues and with eigenvectors and .
- 2
, , .
- 3
, with .
If , what is ?
Symmetric matrices
A symmetric matrix has real eigenvalues and perpendicular eigenvectors, so it is diagonalised by an orthogonal matrix.
A real symmetric matrix () always has real eigenvalues, and eigenvectors for distinct eigenvalues are perpendicular.
If the eigenvectors are normalised, (with them as columns) is an orthogonal matrix: , so . Then .
Perpendicular columns are not enough: each must also have length .
Worked example
has eigenvalues and . Find with .
- 1
gives : . gives : .
- 2
Scalar product: , so they are perpendicular.
- 3
Each has length , so divide by .
A symmetric matrix has an eigenvector for one eigenvalue. Which could be an eigenvector for a different eigenvalue?
When diagonalisation fails
A matrix needs two independent eigenvectors; without them is not invertible.
A matrix with a repeated eigenvalue that has only one eigenvector direction cannot be diagonalised, since would not be invertible.
For example has twice and only the eigenvector .
With complex eigenvalues a real does not exist, though a complex one does. Always check the two independent eigenvectors are not multiples; is a quick test.
Diagonalisable
- Two distinct real eigenvalues
- Symmetric matrices
- Repeated eigenvalue with two eigenvectors, e.g.
Not diagonalisable
- Repeated eigenvalue, one direction, e.g.
- Complex eigenvalues (no real P)
has eigenvalue twice and only one eigenvector direction. Is it diagonalisable?
The Cayley-Hamilton theorem
Every square matrix satisfies its own characteristic equation.
The Cayley-Hamilton theorem: for a matrix with characteristic equation (trace , determinant ),
The constant term becomes , not just . This reduces any power of to a combination of and .
The same holds for matrices, with a cubic.
Worked example
Use Cayley-Hamilton to find for .
- 1
Trace , determinant : .
- 2
So , giving .
A matrix has characteristic equation . Which expresses ?
Inverses and powers from Cayley-Hamilton
Rearrange the relation to for an inverse, or multiply by for higher powers.
Rearrange as . Then .
For higher powers multiply the relation by and substitute repeatedly: .
These give the same results as diagonalisation, so use whichever the question hints at.
- 1
Start from the relation
C² − 6C + 5I = 0
- 2
Factor out C
C(6I − C) = 5I
- 3
Divide by 5 for the inverse
C⁻¹ = (6I − C)/5
- 4
Multiply by C for powers
C³ = 6C² − 5C = 31C − 30I
A matrix satisfies . What is ?
Try an exam question
The matrix . Use the Cayley-Hamilton theorem to find .
[5 marks]
- [1]Characteristic equation .
- [1]Cayley-Hamilton: .
- [1]Rearranges to .
- [1].
- [1].
That's the notes covered.
Carry on to the next subtopic.
Try a question
Which is ?
1 mark, marked against the mark scheme
Which is ?
1 mark, marked against the mark scheme