TheoremBase

Radial Retraction onto a Closed Ball in ℝⁿ and in the Real Matrices

For R > 0, the radial retraction that fixes the points of norm at most R and rescales the others to norm R is a well-defined map into the closed ball of radius R and is Lipschitz with constant 2, both on Euclidean space with the Euclidean norm and on real matrices with the operator norm.

Statement

In the setting of Euclidean Space and Real Matrices: Standing Notation, let n,p,q∈Nn,p,q\in\mathbb{N} and let R∈RR\in\mathbb{R} with R>0R>0. ∥⋅∥\lVert\cdot\rVert is the Euclidean norm on Rn\mathbb{R}^{n} and the operator norm on Mp×q(R)\mathcal{M}_{p\times q}(\mathbb{R}).

There is exactly one map πR:Rn→Rn\pi_{R}:\mathbb{R}^{n}\to\mathbb{R}^{n} such that, for every x∈Rnx\in\mathbb{R}^{n},

πR(x)=xif ∥x∥≤R,πR(x)=R∥x∥ xif ∥x∥>R.\pi_{R}(x)=x\quad\text{if }\lVert x\rVert\le R,\qquad\pi_{R}(x)=\frac{R}{\lVert x\rVert}\,x\quad\text{if }\lVert x\rVert>R.

The same holds for Mp×q(R)\mathcal{M}_{p\times q}(\mathbb{R}) in place of Rn\mathbb{R}^{n}; that map Mp×q(R)→Mp×q(R)\mathcal{M}_{p\times q}(\mathbb{R})\to\mathcal{M}_{p\times q}(\mathbb{R}) is written ΠR\Pi_{R}.

For every x∈Rnx\in\mathbb{R}^{n}, ∥πR(x)∥≤R\lVert\pi_{R}(x)\rVert\le R; and, as in (i), πR(x)=x\pi_{R}(x)=x if ∥x∥≤R\lVert x\rVert\le R. The same holds for ΠR\Pi_{R} and every A∈Mp×q(R)A\in\mathcal{M}_{p\times q}(\mathbb{R}) in place of πR\pi_{R} and xx.

For all x,y∈Rnx,y\in\mathbb{R}^{n},

∥πR(x)−πR(y)∥≤2 ∥x−y∥.\lVert\pi_{R}(x)-\pi_{R}(y)\rVert\le2\,\lVert x-y\rVert.

The same holds for ΠR\Pi_{R} and all A,B∈Mp×q(R)A,B\in\mathcal{M}_{p\times q}(\mathbb{R}) in place of πR\pi_{R}, xx and yy.

Proofs

Log in to submit a proof.

Loading...

Citations

Loading…

Dependencies

Loading…

Related

0 relations

Curated associations between results. These are editable and subjective — they do not replace the dependency graph, which is derived from the references in the text.

No relations recorded yet.

Comments

Log in to comment.

Loading…