The finite section method for infinite Vandermonde matrices (Q692583)

The paper focuses on the solution of the infinite dimensional system of linear equations \(\mathbf{A} x=y\) by the finite section method, where \(\mathbf{A}\) is a infinite Vandermonde matrix \[ \mathbf{A} = \left( \begin{matrix} 1 & 1 & 1&\cdots \\ a_{0} & a_{1} & a_{2}&\cdots \\ a_0^2 & a_1^2 & a_2^2&\cdots \\ \vdots & \vdots & \vdots & \\ \end{matrix} \right). \] The finite section method consists in the idea that the solutions of principal subsystems of growing dimensions \(n\) approximate the solution of the infinite system. The authors define a vector space \(\omega\) of all complex valued sequences and a subspace \(l_1(\alpha)\subset\omega,\) be \(l_1(\alpha)=\{x\in \omega | \sum_r |x_r|\alpha^r<\infty\},\) for real positive \(\alpha\) with the norm given by \(\sum_r |x_r|\alpha^r\). It is shown that if the sequence \(a_0,a_1,a_2,\ldots\) of the coefficients of the Vandermonde matrix \(\mathbf{A}\) meets general convergence conditions the finite section method converges in \(l_1\) sense under condition that the right hand side of the equation is in a suitably weighted \(l_1 (\alpha)\) space.