Maths WorkshopMaths problems with full solutions
Exercise 1772Linear algebra › Matrices and linear systemsL1Added 3 Nov 2026

Orthogonal projection onto a subspace

#orthogonal projection#projection matrix#orthonormal basis

Statement

Let V=span⁡((1,1,0),(0,1,1))V = \operatorname{span}((1,1,0),(0,1,1)) in R3\mathbb{R}^3. Compute the matrix of the orthogonal projection onto VV.

Hint

Build an orthonormal basis of VV using Gram-Schmidt, then use P=QQTP = QQ^T where QQ has the orthonormal basis as columns.

Show solutionHide solution

Solution

Orthonormalization. u1=(1,1,0)/2u_1 = (1,1,0)/\sqrt2. The projection of (0,1,1)(0,1,1) onto u1u_1 is ⟨(0,1,1),u1⟩u1=(1/2)⋅(1,1,0)/2=(1/2)(1,1,0)\langle (0,1,1), u_1 \rangle u_1 = (1/\sqrt2) \cdot (1,1,0)/\sqrt2 = (1/2)(1,1,0). The orthogonal part is
(0,1,1)−(1/2)(1,1,0)=(−1/2,1/2,1).(0,1,1) - (1/2)(1,1,0) = (-1/2, 1/2, 1).
Norm: 1/4+1/4+1=3/2\sqrt{1/4+1/4+1} = \sqrt{3/2}. So u2=2/3(−1/2,1/2,1)=(−1/6,1/6,2/3)u_2 = \sqrt{2/3}(-1/2,1/2,1) = (-1/\sqrt6, 1/\sqrt6, \sqrt{2/3}).

Projection matrix. P=u1u1T+u2u2TP = u_1 u_1^T + u_2 u_2^T:
u1u1T=12(110110000),u2u2T=16(1−1−12/2−1112/2−12/212/24).u_1 u_1^T = \frac12\begin{pmatrix} 1 & 1 & 0 \\ 1 & 1 & 0 \\ 0 & 0 & 0 \end{pmatrix}, \qquad u_2 u_2^T = \frac16\begin{pmatrix} 1 & -1 & -\sqrt{12}/2 \\ -1 & 1 & \sqrt{12}/2 \\ -\sqrt{12}/2 & \sqrt{12}/2 & 4 \end{pmatrix}.
After simplifying (and using 12/2=3\sqrt{12}/2 = \sqrt3):
P=16(42−3243−334).P = \frac16\begin{pmatrix} 4 & 2 & -3 \\ 2 & 4 & 3 \\ -3 & 3 & 4 \end{pmatrix}.

Verification. PP is symmetric, P2=PP^2 = P, and PP fixes VV (check on the generators).