Table of Contents
Fetching ...

KMS Inequalities: From Elliptic Operators to Constant Rank

Paul Stephan

TL;DR

This work extends Korn-Maxwell-Sobolev (KMS) inequalities from elliptic operators to operators with constant rank on the kernel of a given differential constraint $\mathscr{A}$. The main contribution is a left-hand side correction term $\Pi_{\mathbb{B}} \mathcal{P}_{\ker(\mathscr{A})}[P]$, yielding $\| P - \Pi_{\mathbb{B}} \mathcal{P}_{\ker(\mathscr{A})}[P] \|_{\dot W^{k-1,p^*}} \lesssim \| \mathscr{A}[P] \|_{\dot W^{k-1,p^*}} + \| \mathbb{B}P \|_{L^p}$ for $1<p<n$, provided $\mathbb{B}$ has constant rank on $\ker(\mathscr{A})$. The authors establish a partial converse and, for $p=1$, introduce the cancelling condition on $\mathbb{B}$ restricted to $\ker(\mathscr{A})$, linking to ellipticity plus cancelling in the optimal limiting case. A schematic proof builds on Fonseca–Müller, splitting $P$ into kernel and orthogonal components, combined with constant-rank and injectivity arguments, and culminates in domain-related considerations via $\mathbb{C}$-ellipticity. Overall, the paper provides a unified framework to obtain coercive estimates for a broad class of differential operators, including curl/divergence-type structures, with clear implications for variational methods and boundary-domain problems.

Abstract

Korn-Maxwell-Sobolev (KMS) inequalities represent a tool for estimating differential expressions and have gained particular importance in recent years, especially concerning elliptic operators. In my Master's thesis, together with Peter Lewintan (University of Duisburg-Essen), we extended this concept to also apply to operators of constant rank. This makes it possible to cover more complex structures such as the curl or divergence of vector fields. A key difference from the elliptic theory is that in the constant rank case, a correction term $Π_\mathbb{B}$ is necessary on the left-hand side of the inequality. Results were also obtained for the limiting case $p=1$, although additional assumptions are required here. This article provides an illustrative introduction to KMS inequalities and demonstrates their application in both the elliptic and constant rank cases.

KMS Inequalities: From Elliptic Operators to Constant Rank

TL;DR

This work extends Korn-Maxwell-Sobolev (KMS) inequalities from elliptic operators to operators with constant rank on the kernel of a given differential constraint . The main contribution is a left-hand side correction term , yielding for , provided has constant rank on . The authors establish a partial converse and, for , introduce the cancelling condition on restricted to , linking to ellipticity plus cancelling in the optimal limiting case. A schematic proof builds on Fonseca–Müller, splitting into kernel and orthogonal components, combined with constant-rank and injectivity arguments, and culminates in domain-related considerations via -ellipticity. Overall, the paper provides a unified framework to obtain coercive estimates for a broad class of differential operators, including curl/divergence-type structures, with clear implications for variational methods and boundary-domain problems.

Abstract

Korn-Maxwell-Sobolev (KMS) inequalities represent a tool for estimating differential expressions and have gained particular importance in recent years, especially concerning elliptic operators. In my Master's thesis, together with Peter Lewintan (University of Duisburg-Essen), we extended this concept to also apply to operators of constant rank. This makes it possible to cover more complex structures such as the curl or divergence of vector fields. A key difference from the elliptic theory is that in the constant rank case, a correction term is necessary on the left-hand side of the inequality. Results were also obtained for the limiting case , although additional assumptions are required here. This article provides an illustrative introduction to KMS inequalities and demonstrates their application in both the elliptic and constant rank cases.

Paper Structure

This paper contains 8 sections, 42 equations.