27887 Chap.8 Eigenvalues and Eigenvectors
Eigenvalues and Eigenvectors
Directions a matrix does not turn
An eigenvector of a matrix is a non-zero vector that the matrix only stretches, shrinks or reverses, and the factor is its eigenvalue. Eigenvalues come from the characteristic equation, the determinant of A − λI set to zero; each eigenvector comes from solving (A − λI)x = 0, which always has a free constant because the matrix is singular at an eigenvalue.
The 2026 specimen also asks you to recognise an eigenvector directly, by noticing that a matrix sends a vector to a multiple of itself.
Why engineers care
A past example question models masses joined by springs, assumes each mass oscillates, and replaces ω² by λ, so that eigenvalues give the natural frequencies and eigenvectors give the pattern of motion.
Feedback on earlier papers records eigenvectors without a scaling constant, elimination applied to A instead of A − λI, and cubic equations solved when a common factor in the determinant would have left a quadratic.
Where it is tested
Eigenvalues appear in the title of the sixth quiz and close the 2026 specimen, where the last two parts of the matrix question ask you to explain why a vector is an eigenvector and then to find the eigenvector for the second eigenvalue.
Mock papers place them after an inverse in the same question, so an earlier mistake can carry forward unless each part is checked.
What this chapter covers
- 01
The eigenvalue equation Ax = λx
- 02
Characteristic equations from determinants
- 03
Eigenvectors by Gaussian elimination
- 04
Recognising an eigenvector without calculation
- 05
Common factors that avoid a cubic
- 06
Natural frequencies and mode shapes
Worked example · free
Natural modes of two coupled masses
- 2Characteristic equation: (2 − λ)² − 1 = 0, so 2 − λ = ±1 and λ = 1 or λ = 3.
- 1For λ = 1, K − I has rows (1, −1) and (−1, 1), so x = y and the eigenvector is t(1, 1).
- 1For λ = 3, K − 3I has rows (−1, −1) and (−1, −1), so x = −y and the eigenvector is s(1, −1).
- 1Angular frequencies: √1 = 1 and √3.
Key terms
- Eigenvalue
- A scalar λ for which Ax = λx has a non-zero solution x.
- Characteristic Equation
- The polynomial equation det(A − λI) = 0 whose roots are the eigenvalues of A.
- Mode Shape
- The pattern of relative motion described by an eigenvector in a vibrating system.
- Natural Frequency
- A frequency at which a system oscillates freely, given here by the square root of an eigenvalue.
- Scaling Constant
- The free multiplier in an eigenvector, showing that every non-zero multiple is also an eigenvector.
- Trace
- The sum of the diagonal entries of a square matrix, which equals the sum of its eigenvalues.
Eigenvalues and Eigenvectors FAQ
Why must an eigenvector include a free constant?
Any non-zero multiple of an eigenvector is also an eigenvector, so the answer is a direction. Feedback on an earlier paper reports most candidates losing marks for leaving out the scaling constant.
Do I always have to solve a cubic for a three-by-three matrix?
Not when the determinant contains a common factor. The 2018-19 examiners noted that taking such a factor out left a quadratic, while candidates who expanded fully faced a cubic they did not need.
How can I show a vector is an eigenvector without the characteristic equation?
Multiply the matrix by the vector. If the result is a scalar multiple of the vector, it is an eigenvector and the scalar is the eigenvalue; the 2026 specimen asks for exactly this argument, stated in words.
Exam move
Work one two-by-two and one three-by-three eigenproblem in each session, checking that the eigenvalues add up to the trace and multiply to the determinant. For each eigenvector, write the augmented matrix for A − λI, name your row operations and keep a free constant. Then attach a physical meaning to one result, such as the in-phase and out-of-phase modes of two masses.
For complex eigenvalues, practise the quadratic formula with a negative discriminant until writing the square root as j times a real number becomes automatic, since that is where the earlier rotation-matrix answers lost their imaginary part.
Working through Eigenvalues and Eigenvectors in 27887? Sia is AskSia’s AI Mathematics tutor — ask any 27887 Eigenvalues and Eigenvectors question and get a clear, step-by-step explanation grounded in how 27887 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.