scientific article; zbMATH DE number 1150144

From MaRDI portal
Publication:4388338





Let \[ E(A,B,L_p(R^d))= \sup_{x\in A} \sup_{y\in B} \| x-y\| _{L_p(R^d)}. \] The authors consider the average Kolmogorov width \[ \overline{d}_\sigma(C,L_q(R^d))=\inf_L E(C,L,L_q(R^d)), \] where the infimum is taken over all subspaces \(L \subset L_q(R^d)\) such that the average dimension of \(L\) in \(L_q(R^d)\) is not greater than \(\sigma\) and \(C\) is the unit ball in the anisotropic Besov space \(B^{\mathbf r}_{p\theta} (R^d)\). A theorem is formulated which, in particular, states that \(\overline{d}_\sigma(C,L_p(R^d))\) is equivalent to \(\sigma^{-a+ \frac{1}{p}-\frac{1}{q}}\), where \(a=\left(\sum_{j=1}^d\frac{1}{r_j}\right)^{-1}\).











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4388338)