Constructive decomposition of functions of finite central mean oscillation

From MaRDI portal





If \(f\) is a function on \(R^n\), define the \(L^2\)-sharp function \(M_2^\sharp f\) by \[ M_2^\sharp f(x) = \sup | B| ^{-1} \int_B | f - f_B| ^2 \] where B ranges over all balls centered at \(x\), and \(f_B\) is the mean of \(f\) on \(B\). A function is in BMO if \(M_2^\sharp\) is bounded; a function is in CMO (Central Mean Oscillation) if \(M_2^\sharp(0)\) is bounded. Define the Beurling space \(B^2\) to be the space of all functions \(f\) such that \((| f| ^2)_B\) is bounded for all balls \(B\) centered at the origin. This space relates to CMO as \(L^\infty\) relates to BMO. In particular, the author proves that if \((K_1, ..., K_M)\) are singular integral operators satisfying a certain non-degeneracy condition (the Janson-Uchiyama condition), then every CMO function \(g\) can be decomposed as \(g = \sum_{j=1}^M h_j * K_j\), where the \(h_j\) are in \(B^2\). This is in analogy with the Fefferman-Stein decomposition of BMO. The proof uses wavelet bases, and proceeds constructively by first locating an approximate inverse to the operator \((h_j) \mapsto \sum_{j=1}^M h_j * K_j\).











This page was built for publication: Constructive decomposition of functions of finite central mean oscillation

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4243721)