On the halves of a Riordan array and their antecedents (Q2332377)

A Riordan array \(M=(g(x), f(x))\) is defined by a pair of power series \[g(x)=\sum_{n=0}^{\infty}g_nx^n, g_0\neq 0\text{ and }f(x)=\sum_{n=0}^{\infty}f_nx^n, f_0=0, f_1\neq 0.\] A Riordan group is the set of Riordan arrays with multiplication \[(g(x), f(x))\cdot (u(x), v(x))=(g(x)u(f(x)), v(f(x)))\] and inverses \[(g(x),f(x))^{-1}=(\frac {1}{g(\overline{f}(x))},\overline{f}(x) ),\] where \(f(\overline{f}(x))=x\). The author investigates the relationship of the horizontal half and the vertical half of a Riordan array. Using them, the author also provides a canonical factorization of elements of the associated or Lagrange subgroup of the Riordan group. The vertical half matrix is shown to be an element of the hitting-time group. Finally, the author considers some examples which answer to the following question: when can we express a given Riordan array as the half (either vertical or horizontal) of another Riordan array?