The length of the longest head-run in a model with long range dependence (Q5939306)

Let \(\xi_1,\xi_2,\dots\) be a sequence of i.i.d. random variables such that \(P(\xi_1 =+1) = P(\xi_1 =-1) = 1/2\). Let \(L_n\) denote the length of the longest consecutive block of \(+1\)'s in the sequence \(\xi_1,\dots,\xi_n\). Then the Erdős-Rényi law of large numbers [see \textit{P. Erdős} and \textit{A. Rényi}, J. Anal. Math. 23, 103-111 (1970; Zbl 0225.60015)] implies that \(L_n/\log_2(n)\to 1\) \((n\to\infty)\) a.s. In the present paper an Erdős-Rényi law is obtained for a certain stationary sequence of Bernoulli random variables \(X_0,X_1,\dots\) exhibiting a long-range, positive dependence. The definition of \(X_n\) is as follows. Let \(Y(a)\) \((a\in Z^d)\) be a sequence of i.i.d. random variables such that \(P(Y(0)=1)=P(Y(0)=-1)=1/2\), and let \(S_n=W_1+\cdots+ W_n\) \((n\geq 0)\) where \(W_1,W_2,\dots\) is a sequence of i.i.d. random vectors taking values in \(Z^d\) such that \(P(W_1=e_i)=P(W_1=-e_i)=(2d)^{-1}\) (\(e_i\) denoting the usual unit coordinate vectors in \(Z^d\)). Assume that the sequences \((Y(a))\) and \((W_n)\) are independent. Finally put \(X_n =Y(S_n)\) \((n\geq 0)\) and let \(L_n\) denote the length of the longest consecutive block of \(+1\)'s in the sequence \(X_0,\dots,X_n\). The main result is that \[ \lim_{n\to\infty} \frac{L_n}{c_d(\log(n))^{(d+2)/d}}=1\quad \text{a.s.}\quad (d\geq 1) \] (\(c_d\) denoting a constant depending only on \(d\geq 1\)).