On Large Deviations for Stochastic DifferentialEquations with a Small Diffusion and Averaging (Q4954367)

A system of stochastic differential equations \[ \begin{aligned} &dX^\varepsilon_{t} = f(X^\varepsilon_{t}, Y^\varepsilon_{t}) dt + \varepsilon \sigma_{1} (X^\varepsilon_{t},Y^\varepsilon_{t}) dw^{1}_{t} + \varepsilon\sigma_3(X^\varepsilon_{t},Y^\varepsilon_{t}) dw^3_{t},\quad X^\varepsilon_{0} = x, \\ & dY^\varepsilon_{t} = \frac 1{\varepsilon^2} B( X^\varepsilon_{t},Y^\varepsilon_{t}) dt + \frac 1{\varepsilon}C_1(Y^\varepsilon_{t}) d w^1_{t} + \frac 1{\varepsilon}C_{2}(Y^\varepsilon_{t}) dw^2_{t}, \quad Y^\varepsilon_{0} = y, \end{aligned} \] in \(\mathbb{R}^{d}\times\mathbb{R}^{l}\) is considered, where \((w^1, w^2,w^3)\) is a \(3d\)-dimensional Brownian motion and \(\varepsilon>0\) is a small parameter. Sufficient conditions for the ``slow'' components \(\{X^\varepsilon; \varepsilon>0\}\) to obey a large deviation principle in the space \(C([0,1]; \mathbb{R}^{d})\) are given and the rate functional is identified. The proof is sketched only very briefly.