Marcinkiewicz-summability of multi-dimensional Fourier transforms and Fourier series (Q531900): Difference between revisions

Let \(d\in\mathbb{N}\), \(d\geq 1\), and let \(\theta\) be an even function supported on \([-c,c]\), \(0< c\leq\infty\), continuous on \(\mathbb{R}\) and differentiable on \((0,c)\). Suppose \[ \theta(0)= 1,\;\int^\infty_0(t\vee 1)^d|\theta'(t)|\,dt< \infty,\;\lim_{t\to\infty} t^d\theta(t)= 0. \] The Marcinkiewicz-\(\theta\)-means (M-\(\theta\)-m) of the Fourier transform of a function \(f\in L_p(\mathbb{R}^d\)), \(1\leq p\leq 2\), are defined by \[ \sigma^\theta_T f(x)= \int_{\mathbb{R}}\theta(|v|/T)\widehat f(v) e^{2\pi ixv}dv= \int_{\mathbb{R}^d} f(x- v)K^\theta_T(v)\,dv, \] where the kernel \(K^\theta_T\) is given by \[ K^\theta_T(v)= \int_{\mathbb{R}^d}\theta(|v|/T) e^{2\pi ixv}dv,\quad T> 0. \] Under some other weak conditions on \(\theta\), the author extends the definition of M-\(\theta\)-m to all tempered distributions \(f\in H_p(\mathbb{R}^d)\), \(0< p\leq\infty\), and to all functions \(f\in L_p(\mathbb{R}^d)\), \(1\leq p\leq\infty\), by setting 6\(\sigma^\theta_T f:= f* K^\theta_T\). He studies the behaviour of the maximal operator of the M-\(\theta\)-m in \(H_p(\mathbb{R}^d)\) and obtains as a consequence the a.e. convergence of \(\sigma^\theta_T f\) to \(f\) whenever \(f\in L_p(\mathbb{R}^d)\) for \(1\leq p\leq\infty\). Moreover, he proves that the M-\(\theta\)-m are uniformly bounded on the spaces \(H_p(\mathbb{R}^d)\) and so they converge in norm for all \(d\) such that \(d/(d+ a)< p<\infty\), where \(0< \alpha\leq 1\) depends only on \(\theta\). Analogous results hold for Fourier series of functions and distributions on \(\mathbb{T}^1\). Finally, various classical summation methods such as the Féjer, Cesàro, Weierstrass, Picard, Bessel, De La Vallée Poussin, Rogosinski and Riesz methods are considered as particular cases.