Approximation of the inverse \(g\)-frame operator (Q353983): Difference between revisions

Let \({\mathcal H}\) be a Hilbert space, and let \(\{\Lambda_i\}_{i=1}^{\infty}\) be a \(g\)--frame for \({\mathcal H}\) with \(g\)--frame operator \(S\). For \(n \in N\), let \(S_n\) be the \(g\)-frame operator of \(\{\Lambda_i\}_{i=1}^n\). We say that the projection method works if \( \langle g, S_n^{-1}\Lambda^{\ast}_i\Lambda_i \rangle \longrightarrow \langle g, S^{-1}\Lambda^{\ast}_i\Lambda_i \rangle \;\text{as} \;n \longrightarrow \infty \) for every \(f, g \in {\mathcal H}\). We also say that the strong projection method works, if \( \sum_{i=1}^n|\langle f, S_n^{-1}\Lambda^{\ast}_i\Lambda_i \rangle-\langle f, S^{-1}\Lambda^{\ast}_i\Lambda_i \rangle|^2 \longrightarrow 0 \;\text{as} \;n \longrightarrow \infty \) for every \(f \in {\mathcal H}\). In the paper under review, the authors find a necessary and sufficient condition for the projection method to work. In particular, they show that the projection method works for any conditional \(g\)-Riesz frame. They also find sufficient conditions for the strong projection method to work, and develop an approximation method for the inverse \(g\)-frame operator.