Basic Linear Algebra (8hrs) · Engineering Mathematics - I

Quadratic Form

Watch lectures

Quadratic Form

1. Definition

A quadratic form in n variables is a homogeneous polynomial of degree 2 — every term is of second degree. In matrix form it is written with a symmetric matrix:

Q=XTAX,whereAissymmetric(AT=A)andX=(x1,...,xn)TQ = X^T A X, where A is symmetric (A^T = A) and X = (x_1, ..., x_n)^T

For two variables the general form is Q = ax² + 2hxy + by² with matrix

A=[ah;hb]A = [ a h ; h b ]

and for three variables Q = a₁₁x² + a₂₂y² + a₃₃z² + 2a₁₂xy + 2a₂₃yz + 2a₃₁zx with the symmetric matrix A = [a₁₁ a₁₂ a₃₁ ; a₁₂ a₂₂ a₂₃ ; a₃₁ a₂₃ a₃₃].

2. How to Write the Matrix of a Quadratic Form

  • The coefficient of each square (x², y², z²) goes on the diagonal.
  • A cross term 2axy is split equally: a goes in position (1,2) and a in position (2,1). This equal split is exactly what makes A symmetric — and it is the point of the whole construction.
   Q = 2x² + 3y² + 5z² + 4xy + 6yz - 8zx

        [ 2    2   -4 ]        (4xy → 2,2  :  6yz → 3,3  :  -8zx → -4,-4)
   A =  [ 2    3    3 ]
        [ -4   3    5 ]
   Check: XᵀAX = 2x² + 3y² + 5z² + 2(2xy) + 2(3yz) + 2(-4zx)  ✓

3. Nature (Classification) of a Quadratic Form

The nature is decided by the signs of the eigenvalues of A:

Nature Eigenvalues Meaning
Positive definite all λ > 0 Q > 0 for every non-zero X (a bowl touching the origin)
Positive semi-definite all λ ≥ 0, at least one λ = 0 Q ≥ 0, and Q = 0 for some non-zero X
Negative definite all λ < 0 Q < 0 for every non-zero X
Negative semi-definite all λ ≤ 0, at least one λ = 0 Q ≤ 0
Indefinite λ's of both signs Q takes positive and negative values

Sylvester's criterion gives the same answer without computing eigenvalues, using the leading principal minors D₁ = a₁₁, D₂ = det of the first 2 × 2 corner, D₃ = det A, …:

  • Positive definite ⇔ D₁ > 0, D₂ > 0, D₃ > 0 (all positive);
  • Negative definite ⇔ D₁ < 0, D₂ > 0, D₃ < 0 (signs alternate, starting negative);
  • If all Dᵢ ≠ 0 but the signs follow neither pattern → indefinite;
  • If some Dᵢ = 0, the test is inconclusive for definiteness (the form is then semi-definite or indefinite — decide from the eigenvalues).

4. Rank, Index and Signature

Reduce Q by a non-singular transformation X = PY to the canonical (diagonal) form

Q=λ1y12+λ2y22+...+λnyn2(or,afterscaling,y12+y22+...±yp2...)Q = λ_1 y_1² + λ_2 y_2² + ... + λ_n y_n² (or, after scaling, y_1² + y_2² + ... ± y_p² ...)
Quantity Definition How to get it
Rank r number of non-zero terms in the canonical form = number of non-zero eigenvalues = ρ(A)
Index p number of positive terms = number of positive eigenvalues
Signature s (number of positive terms) − (number of negative terms) s = 2p − r

Rank, index and signature are invariants — they do not change under any non-singular (or orthogonal) change of variables, which is why they characterise the form.

   Q = 2x² + 2y² + 2z² + 2xy + 2yz + 2zx

        [ 2  1  1 ]
   A =  [ 1  2  1 ]     eigenvalues  4, 1, 1   (all positive)
        [ 1  1  2 ]

   ⇒  NATURE : POSITIVE DEFINITE
   ⇒  canonical form (using the orthogonal transformation from the Diagonalization topic,
       Q = [1/√3, 1/√3, 1/√3 ; 1/√2, -1/√2, 0 ; 1/√6, 1/√6, -2/√6] ) is
            4Y₁² + Y₂² + Y₃²
   ⇒  rank r = 3 ,  index p = 3 ,  signature s = 2(3) - 3 = +3
   Sylvester check: D₁ = 2 > 0 , D₂ = |2 1 ; 1 2| = 3 > 0 , D₃ = det A = 4 > 0  ✓

7. Why Quadratic Forms Matter in Engineering

Energy expressions (kinetic energy = ½XᵀMX, potential energy of a spring system), stability of equilibrium in mechanics and control theory (a positive definite energy function means a stable equilibrium — the Lyapunov method), the geometry of conics and quadrics (positive definite ↔ ellipse/ellipsoid, indefinite ↔ hyperbola/hyperboloid), principal moments of inertia, and the statistical notion of covariance being positive semi-definite.

8. Exam Points

  • Define a quadratic form and write the matrix of a given form (symmetric A) — 5 marks, almost certain.
  • Find the nature of a quadratic form using eigenvalues, and using Sylvester's criterion — 5-7 marks, very frequent.
  • Define positive definite, semi-definite, negative definite and indefinite forms with the eigenvalue condition — 5 marks.
  • Define and find rank, index and signature of a quadratic form — 5 marks, frequent.
  • Classify a 3-variable form and state its rank/index/signature — 5-7 marks.
  • Show that the nature of a quadratic form is unchanged under a non-singular transformation — 3-5 marks.
  • Give an engineering application of quadratic forms (energy/stability) — 3 marks.
Official AppAndroid App

PYQDeck for Android

Solve BEU previous-year exam papers anywhere, anytime. Fast search, offline paper viewing, and step-by-step solutions designed specifically for your phone.

← All Engineering Mathematics - I modules