Rates of convergence in a central limit theorem for stochastic processes defined by differential equations with a small parameter (Q1201123)

The authors consider the random ordinary differential equation in \(R^ d\), \(\dot X^ \varepsilon(\tau)=F(X^ \varepsilon(\tau),\tau/\varepsilon)\) subject to \(X^ \varepsilon(0)=x_ 0\), where \(\varepsilon>0\) and \(\{F(x,t,\omega)\), \(t\geq 0\}\) is a strong mixing and non-stationary stochastic process, and shows that one can associate with the above equation a certain nonrandom ``averaged'' ordinary differential equation \(\dot x^ 0(\tau)=\overline F(x^ 0(\tau))\) subject to \(x^ 0(0)=x_ 0\) such that if \(\lim_{\varepsilon\to 0}\sup_{0\leq\tau\leq 1}E| X^ \varepsilon(\tau)-x^ 0(\tau)|=0\), then the family of processes \(\{y^ \varepsilon(\tau)\), \(0\leq\tau\leq 1\}\) where \(Y(\tau)=(X^ \varepsilon(\tau)-x_ 0(\tau))/\sqrt\varepsilon\) converges weakly to a certain limiting Gauss-Markov process \(\{\hat Y^ 0(\tau)\), \(0\leq\tau\leq 1\}\) as \(\varepsilon\to 0\). A rate of convergence \(Y^ \varepsilon(\tau)\) to \(\hat Y^ 0(\tau)\) is established.