Ergodic theorems on mixing homogeneous spaces (Q1069208)

The authors consider locally compact homogeneous spaces (mix-spaces) on which all continuous background-free Gaussian random fields are mixing. Euclidean and Lobachevsky spaces are of this kind; the paper contains some other examples of such spaces. The von Neumann ergodic theorem is formulated for such spaces. One of the results is: Let \(X=G/\Gamma\) be a mix-space, \(\xi\) (\(\cdot)\) be a homogeneous random field on X, m an invariant measure on X, and \(\{A_{\alpha}\}\) be the net of measurable sets such that \(0<m(A_{\alpha})<\infty\), \(\alpha\in A\) and \(\lim_{\alpha}m(A_{\alpha})=\infty\). Then \[ (L^ 2)\lim_{\alpha \in A}[m(A_{\alpha})]^{-1}\int_{A_{\alpha}}\xi (x)m(dx)=M(\xi), \] where \(M(\xi)\in L^ 2(\Omega,{\mathcal F},P)\) is the unique mean value such that \(M(\xi)=\Pr_{I_{\xi}}\xi (x)\), \(x\in X\).