Operator estimates using the sharp function (Q1103807)

Let \(f^{\#}\) be the sharp function introduced by Fefferman and Stein. Suppose that T and U are operators acting on the space of Schwartz functions which satisfy the pointwise estimate \((Tf)^{\#}(x)\leq A| Uf(x)|.\) Then, on the \(L^ p\) spaces, the operator norm of T divided by the operator norm of U is bounded by a constant times p. This result allows us to obtain the best possible rate of growth estimate, as \(p\to \infty\), on the norms of singular integral, multiplier, and pseudo- differential operators. The technique also applies to square functions, integral operators with highly oscillating kernels, and fractional integrals. These estimates remain valid on weighted \(L^ p\) spaces defined by an \(A_{\infty}\) weight. The results above are based on an improved good-\(\lambda\) inequality for the sharp function: given \(\epsilon >0\) there is a constant B so that for all \(\lambda >0\), \[ w(\{x\in {\mathbb{R}}^ n:| f(x)| >Bf^{\#}(x)+\lambda\})\leq \epsilon w(\{x\in {\mathbb{R}}^ n:| f(x)| >\lambda\}). \] Using this inequality we are able to derive the norm inequality \(\| f\|_{p,w}\leq C\cdot p\| f^{\#}\|_{p,w}\) for any p, \(1\leq p<\infty\), and any \(w\in A_{\infty}\), where C is independent of p.