Quadratic Forms and Their Matrix Representation
What You'll Learn
You'll learn how to express any quadratic expression in multiple variables as a compact matrix product involving a symmetric matrix.
The Core Idea
A quadratic form is an expression where every term has degree 2. For example:
- In two variables:
Q(x, y) = ax² + bxy + cy² - In three variables:
Q(x, y, z) = ax² + by² + cz² + dxy + exz + fyz
The key insight: any quadratic form can be written as x^T A x, where x is a column vector of variables and A is a symmetric matrix.
How It Works
For two variables, if you have Q(x, y) = 3x² + 4xy + 5y², you can write:
x = [x]
[y]
A = [3 2]
[2 5]
Q = x^T A x = [x y] [3 2] [x]
[2 5] [y]
Notice that the off-diagonal elements share the coefficient of the mixed term: the xy coefficient (4) is split equally between positions (1,2) and (2,1) of matrix A, giving us 2 in each spot. This makes A symmetric.
Why Symmetry?
When you expand x^T A x, the (i,j) and (j,i) entries of A both contribute to the x_i·x_j term. To represent a quadratic form uniquely, we use a symmetric matrix where A_ij = A_ji, distributing each mixed-term coefficient equally.
Three-Variable Example
For Q(x, y, z) = 2x² + 3y² + z² + 4xy - 2xz + 6yz:
A = [2 2 -1]
[2 3 3]
[-1 3 1]
Each mixed term's coefficient is split: the 4xy becomes 2 in positions (1,2) and (2,1), and so on.
Key Takeaway: Every quadratic form can be expressed as x^T A x with a unique symmetric matrix A, where diagonal entries are pure-term coefficients and off-diagonals are half the mixed-term coefficients.