Eigenvalues and eigenvectors
In a minute
An of a is a direction the matrix only stretches: , where the number is its . When a matrix has two different real eigenvalues, they turn any power into a formula in . That is how you predict where a population moving between two towns ends up, years ahead.
You'll be able to
- Test whether a vector is an eigenvector of a matrix, and find its eigenvalue.
- Find the of a matrix and solve it for the eigenvalues, by hand or with a GDC.
- Find an eigenvector for each eigenvalue of a matrix.
- Diagonalize a matrix with two different real eigenvalues as .
- Use to find powers and long-term behaviour, and interpret them in context.
Before you start3 quick questions on what this builds on
Revise first: 1.8 Solving equations with technology and 1.14 Matrices, determinants and inverses
Eigenvalues and eigenvectors
In 1.14 you multiplied a matrix by a column vector. Usually the result points in a new direction. For a few special vectors it stays on the same line: the matrix only stretches them.
Take . Then , which has turned. But : the same direction, three times as long.
So is an eigenvector of with eigenvalue . Two facts follow from .
1. Any non-zero multiple of an eigenvector is also an eigenvector, with the same eigenvalue. and work just as well as .
2. The zero vector never counts, even though for every . An eigenvalue, though, can be or negative. A negative eigenvalue flips the vector to point the opposite way.
To test a vector, work out and see whether it is a multiple of .
Try it first, then reveal one step at a time.
- Multiply: .
- Compare with : . So , and is an eigenvector with eigenvalue .
- Test : .
- No single number times gives : the top entries need the multiplier , the bottom ones . So is not an eigenvector.
Try it first, then reveal one step at a time.
- Read the matrix by columns: column 1 says where Northby's people go ( stay, move). Column 2 does the same for Southam's.
- .
- So : an eigenvector with eigenvalue .
- Interpret: if the populations are in the ratio , they stay exactly the same the next year, and every year after. The leaving the smaller Northby balances the leaving the larger Southam.
The characteristic polynomial
Guessing eigenvectors is slow. Find the eigenvalues first. Write for the , so that .
Then rearranges to . If had an inverse, multiplying by it would give . An eigenvector is never , so has no inverse: its determinant is .
Subtract from each entry on the leading diagonal, then use :
This is not in the formula booklet, but it is quick to rebuild. A matrix has at most two eigenvalues. You are only expected to find eigenvalues, by hand or with technology, for matrices.
Try it first, then reveal one step at a time.
- Subtract on the leading diagonal: .
- Determinant: .
- So the characteristic polynomial is . Check with the formula: and .
- Hence solve : , so the eigenvalues are and .
Try it first, then reveal one step at a time.
- Write down the equation you solve: .
- Expand: .
- GDC polynomial solver with coefficients , , : or
- Interpret: suppose the juveniles and adults are in the proportions of the eigenvector for Then each year the matrix multiplies both groups by The population grows by about a year and keeps those proportions.
Finding the eigenvectors
Once you have an eigenvalue , find an eigenvector by solving with .
1. Write out with the value of put in.
2. Write the two equations from . Because , they say the same thing: one is a multiple of the other.
3. Use an equation that isn't to write in terms of , or in terms of . Give the free one a simple non-zero value, such as , or whatever makes both entries whole numbers.
4. Check that .
Try it first, then reveal one step at a time.
- : , giving and .
- Both say . Take : .
- : , giving , so . Take : .
- Check: and .
Try it first, then reveal one step at a time.
- , giving . The second equation is the same multiplied by .
- So , or . Take , : .
- Every multiple is also unchanged by . The total must be : , so .
- Interpret: with bikes at A and at B, the bikes leaving A each day exactly match the arriving from B. The numbers stay put.
You meet this balanced state again in 4.19, as the steady state of a Markov chain.
Diagonalizing a 2 × 2 matrix
A diagonal matrix has s everywhere off the leading diagonal. Diagonal matrices are easy to work with: to multiply two of them, or to raise one to a power, you just work with the diagonal entries.
To diagonalize , put its eigenvectors side by side as the columns of a matrix . Put the matching eigenvalues on the diagonal of , in the same order:
You only diagonalize matrices with two different real eigenvalues. is not unique: you may multiply either eigenvector by a non-zero number, or swap the columns. If you swap the columns of , swap the eigenvalues in too.
Try it first, then reveal one step at a time.
- Eigenvalue goes with , and eigenvalue with .
- Columns in that order: and .
- , so .
- . Then , which is .
Try it first, then reveal one step at a time.
- and , so the characteristic equation is .
- Solve by factorising or with your GDC: , so or .
- For you already have . For : , so and .
- So and .
- Interpret: moves people from one town to the other with the total fixed. Each year multiplies that part of the populations by . So any gap between the populations and the split shrinks by a year.
Powers of a matrix
In a model , the state after steps is . A GDC finds for you. To see what happens for every , or as grows, you need a formula for .
Diagonalizing gives one. In , the middle is , so . The same happens for every power:
This is in the formula booklet, with the matrix of eigenvectors and the diagonal matrix of eigenvalues. is easy: raise each diagonal entry to the power . Then multiply out the three matrices.
Long-term behaviour comes from the eigenvalues. As , if . If , then for every . If , grows in size without limit, changing sign each step when .
Try it first, then reveal one step at a time.
- .
- .
- . Check: gives , and gives .
- Populations: thousand.
- Long term: , so Northby tends to thousand people and Southam to thousand. The total stays thousand, split as in the eigenvector for .
- Comment: the model assumes the same moving rates every year, with no births, deaths or other migration. That is unlikely to hold for decades, so the short-term figures are more reliable.
Try it first, then reveal one step at a time.
- Read the model: with no rabbits, the foxes fall to each year. With no foxes, the rabbits grow by . The is rabbits eaten by foxes.
- and , so . Then or .
- : gives . : gives . So , and .
- Start with . Then .
- : hundred, about foxes.
- Long term: , so and both get close to . Both populations grow by about a year, in the ratio of the eigenvector for .
- Comment: growth cannot go on for ever on an island with limited space and food, so the model only suits the short term.
Common mistakes
- Putting the eigenvalues in in a different order from their eigenvectors in . Column of must be the eigenvector of the th diagonal entry of .
- Writing . With eigenvectors as the columns of , it is : first, last.
- Giving the zero vector, or no vector, as an eigenvector. Choose a non-zero solution of and check that .
- Rounding an eigenvalue before raising it to a power. Small errors grow when the power is large, so keep the GDC's full value.
- Stopping at or the eigenvalues when the question asks about the context. Say what happens to the quantities, with units, and comment on whether the model is realistic.
Exam tips
- Describe the long-term behaviour: say which terms tend to because , give the limiting values in context with units, and comment on whether the model stays realistic.
- Find an eigenvector: any non-zero multiple of a correct eigenvector earns the marks, so choose whole numbers. Show the equation you get from .
- Write down : put the eigenvalues in the same order as their eigenvectors in the columns of . If is given, first read which eigenvector is in each column.
- Show that is an eigenvector: work out and write it as . Work forwards from and ; don't start from the answer.
Key takeaways
Practise this
26 past-paper questions on Eigenvalues and eigenvectors in the AI HL bank, each with a mark scheme and a worked explanation.