Solving Yosida inclusion problem in Hadamard manifold (Q2189330): Difference between revisions

By reading the first words of the title to be referred to, one recalls the analysis in connection with maximal monotone set-valued operators, zeros of them, resolvents (and fixed points belonging to them), inclusions, Yosida approximations together with the associated convergence statements, and all this, for example, in Hilbert spaces. But the rest of the title reveals that the author uses manifolds to study set-valued inclusions with the above techniques. In the introduction, the author refers to already existing results (especially the use of Korpelevich-type algorithm for inequalities) and to their applicability. Then the task is specified: Find \(x\in K\) such that \[ 0 \in J_{\lambda}^B(x) + B(x), \] where \(K\) is a nonempty closed and convex subset of a finite-dimensional Hadamard manifold (with constant curvature) \(\mathbb M \), where \(B: \mathbb M \Rightarrow \mathbb M\) is a set-valued monotone vector field and \(J_{\lambda}^B\) is the Yosida approximation operator of \(B\). The second chapter contains preparatory notations, definitions, lemmas and propositions as f.i. with respect to the Riemannian metric, projection onto \(K\), geodesics, triangles, monotone vector fields, complementary vector fields, Brouwer's fixed point theorem in the setting of Hadamard manifolds, Fejer convergence, completeness of \(\mathbb M\), the Hopf-Rinow theorem. Chapter 3 is devoted to the main results. There are three steps: First, it is shown, that the given task has a solution.This follows with help of compactness of \(K\), maximal monotonicity of \(B\) and fixed points of the resolvent operator. Second, an algorithm is given (using essentially resolvent properties and projection) and shown that it is well defined. It delivers a sequence \(\{x_k\}\). Third, the convergence of the approximating sequence against a solution of the given inclusion is proved (using especially the Riemannian metric). Chapter 4 (Application) gives an example and it is explained that all the assumptions made for the existence of a solution of an inclusion \(0 \in F(x)+ N_K(x)\) are fulfilled. In Chapter 5, it is mentioned that the proved results extend different similar existing results.