Coincidence point theorems for noncommuting maps (Q972098)

A \(p\)-norm on a linear space \(X\) is a real-valued function \(\|\cdot\|_p\), \(0< p\leq 1\), satisfying (a) \(\| x\|_p\geq 0\) and \(\| x\|_p= 0\iff x= 0\), (b) \(\| x+y\|_p\leq\| x\|_p\), (c) \(\|\lambda x\|_p= |\lambda|^p\| x\|_p\) for all \(x,y\in X\) and all scalars \(\lambda\). The pair \((X,\|.\|_p)\) is called a \(p\)-normed space. For a non-empty subset \(M\) of the space \(X\), let \(CB(M)\) denote the family of non-empty closed bounded subsets of \(M\), let \(H_p\) denote the Hausdorff metric on \(CB(M)\) induced by the metric \(d_p\), i.e. \(H_p(A,B)= \max\{\sup_{x\in A} \text{inf}_{y\in B}\| x- y\|_p\), \(\sup_{x\in B} \text{inf}_{y\in A}\| x- y\|_p\}\), and let \(T: M\to CB(M)\) be a multivalued map. A point \(x\in M\) is a fixed point of \(T\) if \(x\in Tx\). Let the set of fixed points of \(T\) be denoted by \(F(T)\) and the set of coincident points of \(f\) and \(T\) be denoted by \(C(f,T)\). For a single valued mapping \(f: M\to M\) and a multivalued mapping \(T: M\to CB(M)\), the pair \(\{f,T\}\) is called \(R\)-subweakly commuting if for all \(x\in M\), \(fT\in CB(M)\) and there exists \(R> 0\) such that \(M_p(fTx,Tfx)\leq R\text{\,dist}_p(fx,A_\lambda x)\) for every \(0\leq\lambda\leq 1\), where \(A_\lambda z=\lambda Tx+ (1-\lambda)p\) and \(\text{dist}_p(fx, A_\lambda x)= \text{inf}\{\| fx- y_\lambda\|_p: y_\lambda\in A_\lambda x\}\). In this paper, the authors prove some coincidence point theorems for the generalized \((f,g)\)-nonexpansive \(R\)-subweakly commuting multivalued mappings in the setting of \(p\)-normed spaces and also obtain common fixed point thereby generalizing well-known results for noncommuting maps.