On the compactness of commutators of Hardy-Littlewood maximal operator (Q2322524)

Given a collection of locally integrable functions \((b_1,\dots, b_n)\) and a vector function \(\vec{f}=(f_1,\dots,f_m)\) one defines maximal commutators \[\mathcal{M}_{b_i}^{(i)}(\vec{f})(x)=\sup_B \frac{1}{\vert B\vert}\iint_B \vert b_i(x)-b_i(y_j)\vert \prod_{j=1}^m f(y_j) \, dy_1\dots dy_m \, .\] One then defines \[\mathcal{M}_{\Sigma \vec{b}} (\vec{f})(x)=\sum_{i=1}^m \mathcal{M}_{b_i}^{(i)}(\vec{f})(x) \, \] and \[\mathcal{M}_{\prod \vec{b}} (\vec{f})(x) =\sup_{B}\frac{1}{\vert B\vert}\prod_{i=1}^m \int_B \vert b_i(x)-b_i(y_i)\vert \vert f_i(y_i)\vert \, dy_i \, .\] A bi-sublinear operator \(S:X\times Y\to Z\) is called compact if \(S(B_{1,X}\times B_{1,Y})\) is precompact in \(Z\) where \(B_{r,X}=\{x\in X:\Vert x\Vert \leq r\}\). Let \(\text{CMO}\) denote the closure of \(C^\infty\)-functions in the BMO norm.\par The first main theorems establish that \(\mathcal{M}_{\Sigma \vec{b}} \) and \(\mathcal{M}_{\prod \vec{b}} \) are compact operators from \(L^{p_1}\times L^{p_2}\) to \(L^p\) (\(1/p=1/p_1+1/p_2\)), \(1<p_1,p_2<\infty\), provided \(\vec{b}=(b_1,b_2)\) with \(b_1,b_2\in \text{CMO}\).\par Under the same conditions, the same operators are shown to be compact from \(M_{p_1}^{q_1}\times M_{p_2}^{q_2}\to M_p^q\) where, in addition, \(p_i\leq q_i\) and \(1/q=1/q_1+1/q_2\). Here, \(M_p^q\) is the Morrey space defined by \(\Vert f\Vert_{p}^q=\sup_B \vert B\vert ^{1/q}\Bigl(\frac{1}{\vert B\vert}\int_B \vert f(x)\vert ^p\, dx \Bigr)^{1/p}\). In prior work the authors proved that \(L^{p_1}\times\dots \times L^{p_m}\to L^p\)-boundedness of \(\mathcal{M}_{\prod \vec{b}}\) is equivalent to \(b_i\in \text{BMO}\), \(i=1,\dots, m\).