Euclidean projection on a set

An Euclidean projection of a point x_0 \in \mathbb{R}^n on a set {\bf S} \subseteq \mathbb{R}^n is a point that achieves the smallest Euclidean distance from x_0 to the set. That is, it is any solution to the optimization problem

\min\limits_x ||x-x_0||_2: x \in {\bf S}.

When the set {\bf S} is convex, there is a unique solution to the above problem. In particular, the projection on an affine subspace is unique.

Example: assume that {\bf S} is the hyperplane

{\bf S} = \{x\in \mathbb{R}^3: 2x_1+x_2-x_3 = 1\}.

The projection problem reads as a linearly constrained least-squares problem, of particularly simple form:

\min\limits_x ||x||_2 : 2x_1+x_2-x_3 = 1.

The projection of x_0 = 0 on {\bf S} turns out to be aligned with the coefficient vector a=(2,1,-1). Indeed, components of x orthogonal to a don’t appear in the constraint, and only increase the objective value. Setting x=ta in the equation defining the hyperplane and solving for the scalar t we obtain t=1/(a^Ta) = 1/6, so that the projection is x^* = a/(a^Ta) = (1/3, 1/6, -1/6).

License

Hyper-Textbook: Optimization Models and Applications Copyright © by L. El Ghaoui. All Rights Reserved.

Share This Book