Forward-backward and Tseng's type penalty schemes for monotone inclusion problems (Q742190): Difference between revisions

Let \(\mathcal H\) be a Hilbert space and \(M: {\mathcal H} \rightrightarrows {\mathcal H}\) be a set-valued operator. Define \(\mathrm{Gr}(M):=\{(x,u) \in {\mathcal H} \times {\mathcal H}: u \in M(x)\}\) to the graph of \(M\). We say that \(M\) is \textit{monotone} if \(\langle x-y,u-v \rangle \geq 0\) for all \((x,u), (y,v) \in \mathrm{Gr}(M)\). A monotone operator \(M\) is said to be \textit{maximally monotone} if there is no proper monotone extension of the graph of \(M\) on \({\mathcal H} \times {\mathcal H}\). An operator \(D: {\mathcal H} \rightarrow {\mathcal H}\) is said to be \textit{coercive} if there exists \(\alpha > 0\) such that \(\langle x-y, Dx-Dy \rangle \geq \alpha {\parallel Dx-Dy \parallel}^2\) for all \((x,y) \in {\mathcal H} \times {\mathcal H}\). For a subset \(S \subseteq {\mathcal H}\) and for \(x \in S\), let \(N_S(x)\) denote the \textit{normal cone} \(\{u \in {\mathcal H}: \langle y-x, u \rangle \leq 0\) for all \(y \in S\}\). The authors of the article under review study monotone inclusion problems of the form \[ 0 \in A(x) + D(x) + N_C(x), \] where \(A: {\mathcal H} \rightrightarrows {\mathcal H}\) is a maximally monotone operator, \(D: {\mathcal H} \rightarrow {\mathcal H}\) is a coercive operator and \(C \subseteq {\mathcal H}\) is a nonempty set of zeros of another coercive operator \(B: {\mathcal H} \rightarrow {\mathcal H}\). They propose a forward-backward penalty algorithm and prove weak ergodic convergence for a sequence of iterates converging weakly to a zero of the inclusion problem.