The Pascal matrix in the multivariate Riordan group (Q2062810)

For \(n\in \mathbb{N}\), the squared matrices in \(\mathcal{M}_{(n+1)\times (n+1)}(\mathbb{N})\) \[L_n =\left(\binom{i}{j}\right)_ {0\leq i,j\leq n}, \quad U_n = L^T_n, \quad S_n= \left(\binom{i+j}{i}\right)_ {0\leq i,j\leq n} \] are the lower-triangular, upper-triangular and symmetric Pascal matrices of order \(n+1\), respectively. For the vectors \(\textbf{k} = (k_1,\ldots, k_n)\), \(\textbf{i} = (i_1,\ldots, i_n)\in Z^n_{\geq 0}\), \(\textbf{i} \leq \textbf{k}\) indicates the partial order on \(Z^n_{\geq 0}\) defined as \(0 \leq i_j \leq k_j\) for \(1 \leq j \leq n\). Denote the generalized binomials as follows: \[\binom{\textbf{k}}{\textbf{i}}=\binom{k_1}{i_1}\ldots \binom{k_n}{i_n}.\] The multivariate Riordan group is defined in terms of \(n+1\) formal power series in \(C[[x_1, \dots, x_n]]\) as a generalization of the classical (univariate) Riordan group, which is defined in terms of two formal power series in \(C[[x]]\). The author presents an alternative generalization of Pascal matrices. As in the classical case, there are three ways to represent Pascal matrices. Let \(R=\{\textbf{k}_1,\dots, \textbf{k}_r\}\) be a finite set of points in \(Z^n_{\geq 0}\), define the \(r \times r\) matrices \[L_R = (l_{\textbf{k}_i,\textbf{k}_j}),~~U_R = L^t_R, ~~S_R = (s_{\textbf{k}_i,\textbf{k}_j}), \] where \(l_{\textbf{k}_i,\textbf{k}_j}= \binom{\textbf{k}_i}{\textbf{k}_j}\), \(s_{\textbf{k}_i,\textbf{k}_j}= \binom{\textbf{k}_i+\textbf{k}_j}{\textbf{k}_j} \) and the rows and columns of the matrices are ordered by a total order \(<T\) on \(Z^n_{\geq 0}\). The author studies the properties of the Pascal matrices and considers their infinite versions. She also proves that the infinite Pascal matrix is an element of the multivariate Riordan group.