Stochastic two-scale convergence and iterated function system (Q2256044)

This is a report on a talk given by the author at the XXV International Seminar on Stability Problems for Stochastic Models in Maiori, Italy in 2005. The aim is to give an account of a two-scale convergence of a sequence of functions \(u^{k}\) integrated against a test function \(\Phi\) and averaged over the cells of a given level of a fractal determined by a random iterated function system (RIFS), each cell of which is indexed by a word \(\sigma\). The fractal generated by the RIFM is denoted \(D\) and the Hausdorff measure on \(D\) is \(\mu^{*}\). The notion of two scale convergence uses the test functions \(\Phi \in L^{2}(\Omega \times O, \mathcal{B}(\Omega) \times C_b(\mathbb{R}^{n}))\), where \(O\) is an open subset of \(X\), \(\mathcal{B}(\Omega)\) are bounded measurable functions, and \(C_b(\mathbb{R}^{d}\) are bounded continuous functions. Then, one says that \(u^{k} \rightarrow u_0\) in the two-scale sense if \[ \lim_{k \rightarrow \infty} \sum_{|\sigma| = k} r_{\sigma}^{s} \int_{\Omega} \int _X u^{k} \Phi\;d\nu\;d\gamma = \int_{\Omega} \int_D \int_X u_0 \Phi\;d\nu\;d\mu^{*}\;d\gamma. \] This paper is short and contains all the basic definitions for a RIFS and for stochastic two-scale convergence, but does not cover any further applications.