On multivalued nonexpansive mappings in \(\mathbb R\)-trees (Q1760783)

Let \(X\) be a metric space. A \textit{geodesic path} joining the points \(x,y \in X\) is an isometry \(c:[0,l] \rightarrow X\) (for some \(l \geq 0\)) such that \(c(0)=x\) and \(c(l)=y\). If there is a geodesic path between every pair of points in \(X\), then \(X\) is called a \textit{geodesic metric space}. A geodesic metric space \(X\) is called a \textit{metric tree} if there is a unique path (denoted by \([x,y]\)) between any two points \(x, y \in X\) and wherein the following implication holds: \([y,x] \cap [x,z] = \{x\} \Longrightarrow [y,x] \cup [x,z] =[y,z]\), \(x,y,z \in X\). Let \(E\) be a nonempty subset of a metric space \(X\) (with the metric \(d\)) and \(T:E \rightarrow 2^E\). Then \(T\) is called \textit{nonexpansive} if \(H(Tx,Ty) \leq d(x,y)\) for all \(x,y \in X\); \textit{quasi-nonexpansive} if \(\mathrm{Fix}(T) \neq \emptyset\) and \(H(Tx,Tp) \leq d(x,p)\) for all \(x \in X\), \(p\in\mathrm{Fix}(T)\) and \textit{\(*\)-nonexpansive} if for each \(x,y \in E\) and \(u_x \in P_{T(x)}(x)\), there exists \(u_y \in P_{T(y)}(y)\) such that \(d(u_x,u_y) \leq d(x,y)\). Here, \(H\) stands for the Hausdorff metric on \(2^E\) and \(\mathrm{Fix}(T)\) denotes the set of all fixed points of \(T\) (\(x \in E\) is called a \textit{fixed point} of \(T\) if \(x \in T(x)\)). \(P_{T(x)}\) denotes the metric projection onto \(T(x)\). In the paper under review, the authors study relationships between nonexpansive, quasi-nonexpansive and \(*\)-nonexpansive maps, among others, in complete metric trees. Let us state a sample result. If \(E\) is a nonempty subset of a complete metric tree \(X\) and \(T\) is a multi-valued mapping taking closed and convex values in \(2^E\), then \(T\) is weakly nonexpansive if and only if \(T\) is nonexpansive. Convergence of certain iterative processes to a fixed point of \(T\) are also proved in the presence of some sufficient conditions.