A double-sequence random iteration process for random fixed points of contractive type random operators (Q1422288)

Let \(C\) be a closed convex subset of a Hilbert space, \((\Omega ,\Sigma )\) a measurable space, and \(T:\Omega\times C\to C\) a continuous random operator. The author considers the iteration scheme \[ f_{k,n+1}(\omega )=(1-c_{n})f_{k,n}(\omega )+c_{n}T_{k}(\omega ,f_{k,n} (\omega )), \] where \(T_{k}(\omega ,x)=(1-b_{k})t+b_{k}T(\omega ,x)\), \(t\in C\) is fixed, \(c_{n},b_{k}\in (0,1)\) and \(\lim_{n\to\infty}c_{n}=h\) with \(0<h<1\), \(\lim_{k\to\infty}b_{k}=1\). The main result of this note says that if \(T\) satisfies some contraction-type condition and the double sequence \((f_{k,n})\) converges to \(f\), then \(f\) is a random fixed point of \(T\). Reviewer's remark: It is immediate that if \(0<h<1\) and \((f_{k,n})\) converges, then its limit is a random fixed point of \(T\), and we do not need any contraction conditions.