Subdyadic square functions and applications to weighted harmonic analysis (Q502635)

Let \(\alpha,\,\beta\in\mathbb{R}\) and \(\mathbb{R}^n\) be the \(n\)-dimensional Euclidean space. Assume that \(\phi\) is a smooth function with suitable compact Fourier support away from the origin. For any \(x\in\mathbb{R}^n\) and \(t\in(0,\infty)\), let \(\phi_t(x):=t^{-n}\phi(x/t)\). Moreover, for any \(x\in\mathbb{R}^n\), let \[ \Gamma_\alpha(x):=\{(y,t)\in\mathbb{R}^n\times(0,\infty):\;t^\alpha\leq1,\;|y-x|\leq t^{1-\alpha}\}. \] Then, for suitable function \(f\) on \(\mathbb{R}^n\), the square function \(g_{\alpha,\,\beta}(f)\) is defined by, for any \(x\in\mathbb{R}^n\), \[ g_{\alpha,\,\beta}(f)(x):=\left\{\int_{\Gamma_\alpha(y)}|f\ast\phi_t(y)|^2\frac{dy}{t^{(1-\alpha)n+2\beta}} \frac{dt}{t}\right\}^{1/2}. \] Let \(\lambda\in(0,\infty)\). Then the square function \(g^\ast_{\alpha,\,\beta,\,\lambda}(f)\) is defined by, for any \(x\in\mathbb{R}^n\), \[ g^\ast_{\alpha,\,\beta,\,\lambda}(f)(x):=\left\{\int_{\{t^\alpha\leq1\}} \int_{\mathbb{R}^n}|f\ast\phi_t(y)|^2\left[1+\frac{|x-y|}{t^{1-\alpha}}\right]^{-n\lambda} \frac{dy}{t^{(1-\alpha)n+2\beta}}\frac{dt}{t}\right\}^{1/2}. \] Denote by \(T_m\) the convolution operator with Fourier multiplier \(m\). In this paper, under some assumptions for the Fourier multiplier \(m\) and \(\lambda\), the authors establish the pointwise estimate \[ g_{\alpha,\,\beta}(T_mf)(x)\leq Cg^\ast_{\alpha,\,0,\,\lambda}(f)(x) \] for any \(x\in\mathbb{R}^n\) and suitable function \(f\), where \(C\) is a positive constant independent of \(f\) and \(x\). Furthermore, some weighted \(L^2\) inequalities, relating to \(g_{\alpha,\,\beta}(f)\), \(g^\ast_{\alpha,\,0,\,\lambda}(f)\) and \(T_m(f)\), were also obtained. Moreover, some applications of these results to solution operators for dispersive PDE and oscillatory convolution operators were given.