A nonconventional invariance principle for random fields (Q354752)

Let \(X(n)\), \(n\in\mathbb{Z}^{\nu}\), be a sufficiently well mixing vector-valued random field satisfying some moment conditions and stationarity assumptions. Given positive integers \(k\leq l\), consider the functions \(q_j(n)=jn\), \(j\leq k\), and suppose that \(q_i:\mathbb{Z}^{\nu}\to\mathbb{Z}^{\nu}\), \(i=k+1,\dots,l\), map \(\mathbb{Z}^{\nu}_+\) into itself. For \(N\in\mathbb{N}\) and \(t=(t_1,\dots,t_{\nu})\in [0,1]^{\nu}\) introduce \( \xi_N(t)=N^{-\nu/2} \sum (F(X(q_1(n)),\dots,X(q_l(n)))- \overline{F})\) where the sum is over \(n \in \mathbb{Z}^{\nu}_+\) such that \(n_i\leq Nt_i\), \(i=1,\dots,\nu\). Here, \(F\) is a continuous function with polynomial growth and certain regularity properties, \(\overline{F}= \int Fd(\mu\times\ldots \times \mu)\) where \(\mu\) is the distribution of each \(X(n)\). Assuming some growth conditions of \(|q_i|\), \(i>k\), in \(|n|\) (where \(|\cdot|\) stands for the Euclidean distance) the author proves that the random field \(\xi_N(\cdot)\) converges in distribution (in the Skorokhod space) to the \(l\)-dimensional Gaussian random field on \([0,1]^{\nu}\) having specified covariance structure. Thus an extension of the paper by \textit{Yu. Kifer} and \textit{S. R. S. Varadhan} [``Nonconventional limit theorems in discrete and continuous time via martingales'', Ann. Probab. (to appear)] devoted to a nonconventional invariance principle for stochastic processes is provided. However, for random fields the author has to use another way as the martingale techniques is not appropriate.