Strong approximations of Brownian sheet by uniform transport processes (Q2173248)

Let \(W=\{W(s,t) : (s,t)\in[0,1]^2\}\) be a Brownian sheet, i.e., a zero mean continuous real-valued Gaussian process with covariance function \(\mathbb{E}(W(s_1,t_1)W(s_2,t_2))=\min(s_1,s_2)\min(t_1,t_2)\), \((s_1,t_1), (s_2,t_2)\in[0,1]^2\). Using a family of independent standard Poisson processes and an independent sequence of random variables having Bernoulli distribution with parameter \(\frac{1}{2}\), for any positive integer \(n\) and positive real number \(\lambda\), the authors introduce a stochastic process \(W_n^{(\lambda)}=\{W_n^{(\lambda)}(s,t) : (s,t)\in[0,1]^2\}\) (can be called a transport process) and they prove that there exists realizations of \(W_n^{(\lambda)}\) with \(\lambda\in(0,\frac{1}{5})\) on the same probability space as the Brownian sheet \(W\) such that \[ \lim_{n\to\infty} \max_{0\leq s,t\leq 1} \vert W_n^{(\lambda)}(s,t) - W(s,t)\vert = 0 \text{ almost surely.} \] They also extend this result to the general case of a \(d\)-parameter Brownian motion. The proofs are based on an extension of Skorokhod's result on a reproduction of independent random variables by evaluating a Brownian motion at random times, and on the method of the paper [\textit{R. J. Griego} et al., Ann. Math. Stat. 42, 1129--1131 (1971; Zbl 0216.21405)].