ON L^p–L^q TRACE INEQUALITIES

From MaRDI portal
Publication:3419086




Abstract: We give necessary and sufficient conditions in order that inequalities of the type | T_K f|_{L^q(dmu)}leq C |f|_{L^p(dsigma)}, qquad f in L^p(dsigma), hold for a class of integral operators TKf(x)=intRnK(x,y)f(y)dsigma(y) with nonnegative kernels, and measures dmu and dsigma on Rn, in the case where p>q>0 and p>1. An important model is provided by the dyadic integral operator with kernel KmathcalD(x,y)sumQinmathcalDK(Q)chiQ(x)chiQ(y), where mathcalD=Q is the family of all dyadic cubes in Rn, and K(Q) are arbitrary nonnegative constants associated with QinmathcalD. The corresponding continuous versions are deduced from their dyadic counterparts. In particular, we show that, for the convolution operator Tkf=kstarf with positive radially decreasing kernel k(|xy|), the trace inequality | T_k f|_{L^q(dmu)}leq C |f|_{L^p(d x)}, qquad f in L^p(dx), holds if and only if mathcalWk[mu]inLs(dmu), where s=fracq(p1)pq. Here mathcalWk[mu] is a nonlinear Wolff potential defined by and . Analogous inequalities for 1leq<p were characterized earlier by the authors using a different method which is not applicable when q<1.




Cited in
(26)






This page was built for publication: ON $L^p$–$L^q$ TRACE INEQUALITIES

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3419086)