Representation of bounded solutions of linear discrete equations (Q2202859)

Given a countable set \(G\) and a collection of finite-dimensional Banach spaces \(E_{g}\) (\(g\in G\)), denote by \(\mathcal{M}\) the set of mappings \(x:G\to \bigcup_{g\in G}E_{g}\) such that \(x(g)\in E_{g}\) for all \(g\in G\) and \(\|x\|_{\mathcal{M}}=\sup_{g\in G}\|x(g)\|_{E_{g}}<\infty\). Consider the equation \[ \sum_{\alpha\in G}A(g,\alpha)x(\alpha)=y(g)\in\mathcal{M},\ g\in G,\tag{1} \] where \(A(g,\alpha)\in L(E_{\alpha},E_{g})\). The main result is that, if equation (1) is uniquely solvable for every \(y(g)\in\mathcal{M}\) then there exist linear operators \(\Omega(g,\alpha)\in L(E_{\alpha},E_{g})\) such that a solution \(x(g)\in\mathcal{M}\) to (1) is representable as \[ x(g)=\sum_{\alpha\in G}\Omega(g,\alpha)y(\alpha),\ \ \sup_{g\in G}\sum_{\alpha\in G}\|\Omega(g,\alpha)\|_{L(E_{\alpha},E_{g})}<\infty. \]