The variable exponent Sobolev capacity and quasi-fine properties of Sobolev functions in the case \(p^-=1\) (Q2019089): Difference between revisions

Transferring the well-known properties of the classical Lebesgue spaces to the case of variable exponent spaces often requires a specific attitude, particularly in the cases when the exponent \(p(x)\) is not bounded away from 1 and \(+\infty\). If \(p^-=\inf p>1\) it has been proved that every variable exponent Sobolev function has Lebesgue points outside a set of \(p(\cdot)\)-capacity zero and that the precise pointwise representative of a variable Sobolev function is \(p(\cdot)\)-quasi-continuous -- see [\textit{P. Harjulehto} and \textit{P. Hästö}, Ann. Acad. Sci. Fenn., Math. 29, No. 2, 295--306 (2004; Zbl 1079.46022)]. The authors of the paper consider the case \(p:\mathbb R^n\to[0,+\infty)\) including the case of \(p^-=\inf p\geq1\), \(p^+=\sup p=+\infty\) and prove the subadditivity of the \(p(\cdot)\)-capacity. Since it is not possible to use the reflexivity of the Sobolev space and the boundedness of the maximal operator in the limiting case, they consider instead an operator defined as a limit of integral averages and for this operator they obtain a capacitary weak-type estimate. In this way they prove that variable Sobolev functions have Lebesgue points quasi-everywhere and that these functions have quasi-continuous representatives.