A solution to the Bézout equation in \(A(K)\) without Gelfand theory (Q444116): Difference between revisions

Let \(M\subseteq\mathbb{C}\) and \(K\subseteq M\). Let \(A(K,M)\) denote the uniform algebra of all functions that are continuous on \(M\) and holomorphic on the interior of \(K\). The authors present a constructive proof of Arens' result that for \((f_1,\dots, f_n)\in A(K,M)^n\) the Bézout equation has a solution in \(A(K,M)\) if and only if the functions \(f_j\) have no common zero in \(K\). In particular, they show that given \((f_1,\dots, f_n)\in A(K,M)^n\) there exists \((g_1,\dots, g_n)\in A(K,M)^n\) such that \[ \sum_{j=1}^n f_j(z)g_j(z)=1\quad\forall z\in K \] if and only if the functions \(f_j\) have no common zero in \(K\). The authors also consider the case when the functions \(F_j\) are matrix-valued, i.e, each \(F_j\) is a \(N\times N\) matrix with entries in \(A(K,M)\). In this case, they show that given \(F_1,\dots, F_n\) matrix-valued functions, there exists \(G_1,\dots, G_n\) such that \[ \sum_{j=1}^n G_j(z)F_j(z)=1\quad\forall z\in K \] if and only if all \((N,N)\) minors of the matrix \[ \left( \begin{matrix} F_1\\ \vdots\\ F_n \end{matrix} \right) \] have no common zeros in \(K\).