Multilinear Fourier multipliers with minimal Sobolev regularity. II (Q2360946): Difference between revisions

Let \(m\in\mathbb{N}\). For the Schwartz function \(f\) on \(\mathbb{R}^n\), the Fourier transform \(\widehat{f}\) of \(f\), is defined by \[ \widehat{f}(\xi):=\int_{\mathbb{R}^n}f(x)e^{-2\pi ix\cdot\xi}\,dx. \] Let \(\sigma\) be a bounded function on \(\mathbb{R}^{nm}:=\mathbb{R}^n\times\cdots\times\mathbb{R}^n\). Then, the \(m\)-linear Fourier multiplier operator \(T_\sigma\) is defined by, for any \(x\in\mathbb{R}^n\), \[ T_\sigma(f_1,\ldots,f_n)(x):=\int_{\mathbb{R}^{nm}}e^{2\pi ix\cdot(\xi_1+\cdots+\xi_m)} \sigma(\xi_1,\ldots,\xi_m)\widehat{f}_1(\xi_1)\cdots\widehat{f}_m(\xi_m)\,d\vec{\xi}, \] where \(d\vec{\xi}:=d\xi_1\cdots d\xi_m\). For \(p\in(0,\infty]\), denote by \(H^p(\mathbb{R}^n)\) the Hardy space on \(\mathbb{R}^n\). When \(p\in(1,\infty]\), the Hardy space \(H^p(\mathbb{R}^n)\) is just the Lebesgue space \(L^p(\mathbb{R}^n)\). Let \(p_1,\,p_2,\,\ldots,\,p_m\in(0,\infty]\) and \(p\in(0,\infty)\). Assume further that \[ \frac1{p_1}+\frac1{p_2}+\cdots+\frac1{p_m}=\frac1p. \] In this paper, the authors obtained the optimal condition for the multiplier \(\sigma\), in terms of local \(L^2\)-Sobolev space estimates, such that the \(m\)-linear Fourier multiplier operator \(T_\sigma\) is bounded from \(H^{p_1}(\mathbb{R}^n)\times\cdots\times H^{p_m}(\mathbb{R}^n)\) to \(L^p(\mathbb{R}^n)\). For part I, see [\textit{L. Grafakos} and \textit{H. van Nguyen}, Colloq. Math. 144, No. 1, 1--30 (2016; Zbl 1339.42012)].