Sharp \(L^{p,\infty}\to L^q\) estimates for the dyadic-like maximal operators (Q485268): Difference between revisions

Let \({\mathcal M}_d\) be the dyadic maximal operator on \(\mathbb{R}^n\) defined for \(\phi\), a locally integrable function, by \[ {\mathcal M}_d \phi(x)=\sup \{ \frac{1}{|Q|} \int_Q |\phi(u)| \;du: \;x\in Q \}, \] where the supremum is taken over the dyadic cubes \(Q\) formed by the grids \(2^{-N} \mathbb{Z}^n\), where \(N\geq 0\) is an integer. The paper gives an explicit formula for the best constant in the estimate of the norm of \({\mathcal M}_d\) as an operator from \(L^{p,\infty}(\mathbb{R}^n)\) to \(L^q(\mathbb{R}^n)\), \(1\leq q<p<\infty\). Moreover, the optimal constant is shown to be equal to \(\displaystyle{\left(\frac{p}{p-q}\right)^{1/q} \frac{p}{p-1}}\). Actually, the result is obtained in the more general setting of maximal operators \({\mathcal M}_{\mathcal T}\) associated with a tree-like structure \({\mathcal T}\) of measurable subsets of a non-atomic probabilistic space \((X, \mu)\). The proof is developed making use of the so-called Bellman function method which, roughly speaking, relates the problem of proving a given inequality for the maximal operator to the existence of a special function having certain concavity and majorization properties.