Littlewood-Paley-Stein type square functions based on Laguerre semigroups (Q653822): Difference between revisions

Let \[ L_{\alpha}=2| \alpha| +2d+\sum_{j=1}^d\delta_j^*\delta_j \] (where \(\delta_j=\frac{\partial}{\partial x_j}+x_j,\) \(\delta_j^*=-\frac{\partial}{\partial x_j}+x_j-\frac{2\alpha_j+1}{x_j},\) \(j=1,\ldots ,d)\) be the Laguerre differential operator that plays the role of Laplacian for the space \((\mathbb R_+^d,d\mu_{\alpha},| \cdot | ),\) \(d\mu_{\alpha}(x)=x_1^{2\alpha_1+1}\ldots x_d^{2\alpha_d+1}dx,\) \(x=(x_1,\ldots ,x_d)\in \mathbb R_+^d.\) The Laguerre functions \(\ell_k^{\alpha}(x)=\prod_{i=1}^d\ell_{k_i}^{\alpha_i}(x_i),\) where \(\ell_{k_i}^{\alpha_i}\) are the one-dimensional Laguerre functions, are eigenfunctions of \(L_{\alpha}\) with the corresponding eigenvalues \(4| k| +2| \alpha| +2d=\lambda_{| k| }^{\alpha}.\) The spectral decomposition of a natural self-adjoint extension \(\mathcal L_{\alpha}\) of \(L_{\alpha}\) is given by \(\mathcal L_{\alpha}f=\sum_{n=0}^{\infty}\lambda_n^{\alpha}\mathcal P_n^{\alpha}f,\) where \(\mathcal P_n^{\alpha}f=\sum_{| k| =n}\langle f,\ell_k^{\alpha}\rangle_{d\mu_{\alpha}}\ell_k^{\alpha}.\) The Laguerre heat semigroup and its ``modifications'' are generated by \(\mathcal L_{\alpha}\) and \(\mathcal M^{\alpha}_j\) as \[ T_t^{\alpha}=\exp(-t\mathcal L_{\alpha}),\quad \tilde T_t^{\alpha,j}=\exp(-t\mathcal M_j^{\alpha}),\, t\geq 0, \] where \[ \mathcal M_j^{\alpha}f=\sum_{n=1}^{\infty} \lambda_n^{\alpha}\mathcal P_n^{\alpha,j}f,\quad \mathcal P_n^{\alpha,j} f=\sum_{| k| =n}\langle f,x_j\ell_{k-e_j}^{\alpha+e_j} \rangle_{d\mu_{\alpha}} x_j\ell_{k-e_j}^{\alpha+e_j}, \] and \(e_j\) is the \(j\)th coordinate vector. The Poisson semigroups \(P_t^{\alpha},\tilde P_t^{\alpha,j},\) associated with \(\mathcal L_{\alpha}\) and \(\mathcal M_j^{\alpha},\) respectively, are given by \[ P_t^{\alpha}f=e^{-t\sqrt{\mathcal L_{\alpha}}} f,\quad \tilde P_t^{\alpha,j}f=e^{-t\sqrt{\mathcal M_j^{\alpha}}}f, \quad f\in L^2(d\mu_{\alpha}),\, j=1,\ldots ,d. \] The vertical and horizontal \(g\)-functions based on those semigroups are defined as \[ g_{D,G}(f)(x)=\|\partial_DG_tf(x)\|_{L^2(w_D)}, \] where \(G\) denotes either \(T,\tilde T, P\) or \(\tilde P,\) \(G_t\) is equal to either \(T_t^{\alpha},\tilde T_t^{\alpha,j}, P_t^{\alpha},\) or \(\tilde P_t^{\alpha,j},\) respectively; \(D\) is either \(V\) or \(H,\) \(\partial_D=\partial_t\) for \(D=V,\) \(\partial_H=\delta_i\) or \(\partial_H=\delta_j^*\) depending on \(G\) (the last case corresponds to \(\tilde T_t^{\alpha,j}\) or \(\tilde P_t^{\alpha,j}\) when \(i=j);\) \(w_D=tdt\) when \(D=V\) and \(w_D=dt\) when \(D=H\) for the Laguerre heat semigroup and its modifications, and \(w_D=tdt\) for the Poisson semigroups. The main result: for \(\alpha\in[-1/2,\infty)^d\) each of \(g\)-functions from above, viewed as a vector-valued operator related to \(L^2(w_D),\) is a Calderón-Zygmund operator in the sense of \((\mathbb R_+^d,d\mu_{\alpha},| \cdot | ).\)