Radial Fourier multipliers in high dimensions (Q536616)

The authors study convolution operators with radial kernels acting on functions defined on \(\mathbb R^d\). These can also be described as Fourier multiplier transformations \(T_m\) defined by \(\widehat{T_mf}=m\hat f\) with radial function \(m\). A simple characterization of convolution operators bounded on \(L^p\) (whether radial or not) is known only in two cases: \(p=1\) and \(p=2\). It is currently widely believed that for \(p\in (1,2)\), a full characterization of all \({\mathcal F}L^p\) multipliers in reasonable terms is impossible. In this paper, the authors deal with the class of radial multipliers. By combining the weak orthogonality properties with the support size estimates, the authors establish a simple and effective characterization of the radial multiplier transformations bounded on the entire space \(L^p(\mathbb R^d)\) if the dimension \(d\) is sufficiently large. To be precise, let \(d\in[4,\infty)\), \(1<p<p_d:=\frac{2d-2}{d+1}\), and let \(m\) be radial. In this paper, the authors show that for any fixed \(\eta\in{\mathcal S}(\mathbb R^d)\) which is not identically 0, \[ \|T_m\|_{L^p\mathbb R^d)\to L^p(\mathbb R^d)} \sim\sup_{t>0}t^{\frac{d}{p}}\|T_m[\eta(t\cdot)] \|_{L^p(\mathbb R^d)}. \] As a special situation, when \(m\) is compactly supported away from the origin and \(1<p<p_d\), then the convolution operator \(T_m\) is bounded on \(L^p(\mathbb R^d)\) if and only if the (radial) convolution kernel \(\widehat m\) belongs to \(L^p(\mathbb R^d)\). The authors also state an extension of this result to \(H^p\) spaces for \(p\leq1\), which holds for dimensions \(d\geq2\); moreover, the authors obtain Lorentz space bounds (including weak type \((p,p)\) inequalities). Applying their new method, the authors obtain some improvements upon the existing results in the so-called local smoothing problem for the wave equation in high dimensions which is important in harmonic analysis. Indeed, the authors prove that if \(d\in[4,\infty)\) and \(q>q_d:=2+\frac{4}{d-3}\), then there exists a positive constant \(C_{q,d}\) such that for all \(L>0\), \[ \frac{1}{2L}\int_{-L}^L \|e^{it\sqrt{-\Delta}}f\|_{L^q(\mathbb R^d)}^q\,dt \leq C_{q,d}^q\|(I-L^2\Delta)^{\frac{\alpha}{2}}f\|_{L^q(\mathbb R^d)}^q \] holds for \(\alpha=\alpha(q)=d(\frac{1}{2}-\frac{1}{q})-\frac{1}{2}\) and all suitable functions \(f\), which can be strengthened further by using suitable Triebel-Lizorkin spaces.