The isoperimetric theorem for general integrands (Q1345482): Difference between revisions

\S 1 contains historical and bibliographical notes about isoperimetric problems. \S 2 refines the terminology for sets with finite perimeter in \(\mathbb{R}^n\), \(n\geq 2\). \S 3 defines a positive, constant coefficient, parametric integrand of degree \(n-1\) on \(\mathbb{R}^n\). \S 4 gives a new proof for the following isoperimetric theorem: ``Let \(\psi\) be a norm on \(\mathbb{R}^n\), or, more generally, a positive, constant coefficient, parametric integrand of degree \(n-1\) (not necessarily even or convex). Let \(B_\psi= \{x: \psi^*(x)\leq 1\}\) be the unit ball or the Wulff crystal. Let \(\Omega\) be any measurable subset of \(\mathbb{R}^n\) with finite perimeter and of the same volume as \(B_\psi\). Then \(\psi (\partial\Omega)\geq \psi(\partial B_\psi)\), with equality holding if and only if \(\Omega\) differs from a translate of \(B\psi\) by a set of volume zero.'' \S 5 derives an extension of the Gauss-Green-Federer divergence theorem which is suitable for the purposes of this paper.