Coincidence and invariant approximation theorems for generalized \(f\)-nonexpansive multivalued mappings (Q885623): Difference between revisions

The authors prove the following result. Let \(X\) be a normed space, \(M\subset X\) nonempty and complete, \(q\in M\) and \(f:M\to M\) a surjective map. Let \(T\) be a mapping from \(M\) into the closed nonempty subsets of \(M\) such that \[ H(Tx,Ty)\leq\max\{\| fx-fy\|,\;d(fx,T_\lambda x),\;d(fy,T_\lambda y),\;\tfrac12[d(fx,T_\lambda y)+d(fy,T_\lambda x)]\} \] for each \(x,y\in M\) and \(\lambda\in[0,1]\), where \(T_\lambda x=(1-\lambda)q+\lambda x\) and \(H\) denotes the Hausdorff distance. Assume that \(T(M)\) is bounded and \((f-T)(M)\) is closed. Assume, moreover, that there is a sequence \((k_n)\) in \((0,1)\) converging to 1 such that \(T_{k_n}x\subset M\) for each \(x\in M\) and \(n\in\mathbb{N}\). The conclusion is that there is an \(x\in M\) such that \(fx\in Tx\).