← 🎯 Eigenvalues & Eigenvectors
Subtopic 1 of 6

5.1Eigenvalues, eigenvectors and the characteristic polynomial

The idea

An eigenvector is a direction that a matrix only stretches, never turns: Av = λv. The stretch factor λ is the eigenvalue.

Think of it like this: Spinning a globe: every point moves, except the axis, which stays on its own line.

Why it exists

Eigenvalues summarise a matrix: they give its trace, determinant, powers and long-term behaviour.

How it works

  1. Eigenvalues are the roots of det(A − λI) = 0 (the characteristic equation).
  2. For each λ, solve (A − λI)v = 0 to get the eigenvectors.
  3. An n × n matrix has n eigenvalues counted with multiplicity (possibly complex).
  4. For 2 × 2: λ² − (trace)λ + det = 0.

Worked example: A = [[4, 1],[2, 3]]

λ² − 7λ + 10 = 0, so λ = 2 and 5. The lab gives the eigenvector for each, and checks that the eigenvalues add up to the trace 7 and multiply to the determinant 10.

Now play with it

  • →Load each example and read the characteristic polynomial.
  • →Load the rotation: no real eigenvalue, so no real direction is preserved.
det(A − λI) = λ^2 − 7λ + 10 = 0
eigenvalue λalgebraic mult.geometric mult.eigenvectors (basis of the eigenspace)
211(-1/2, 1)
511(1, 1)
Σ λ vs trace7 vs 7
Π λ vs det10 vs 10
Singular?no
Diagonalisable?yes: geometric = algebraic for every λ

Check yourself

The eigenvalues of [[4, 1],[2, 3]] are

Remember: Av = λv; eigenvalues are the roots of det(A − λI) = 0.
Exam notes (the dense version)

A non-zero vector v is an eigenvector of A with eigenvalue λ if Av = λv: A only stretches v. Eigenvalues are the roots of the characteristic equation det(A − λI) = 0; for each λ, solve (A − λI)v = 0 for the eigenvectors. An n×n matrix has n eigenvalues counted with multiplicity (some may be complex). For 2×2: λ² − (trace)λ + det = 0.

Previous year questions: Eigenvalues & Eigenvectors (14)

PYQs are grouped by topic, so this is the full set for Eigenvalues & Eigenvectors; try the ones that match this subtopic. Open all 14 on the topic page.