Abstract:This work is concerned with the large deviation principle for a family of slow-fast systems perturbed by infinite-dimensional mixed fractional Brownian motion (FBM) with Hurst parameter $H\in(\frac12,1)$. We adopt the weak convergence method which is based on the variational representation formula for infinite-dimensional mixed FBM. To obtain the weak convergence of the controlled systems, we apply the Khasminskii's averaging principle and the time discretization technique. In addition, we drop the boundedness assumption of the drift coefficient of the slow component and the diffusion coefficient of the fast component.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is to establish the **Large Deviation Principle (LDP)** for a class of fast - slow systems driven by infinite - dimensional mixed fractional Brownian motion (FBM). Specifically, the research object is a fast - slow system on the time interval \([0, T]\), where its slow variable \(X^{\varepsilon, \delta}_t\) and fast variable \(Y^{\varepsilon, \delta}_t\) are perturbed by fractional Brownian motion and standard Brownian motion respectively.
### Research Background and Problem Description
1. **System Description**:
This fast - slow system is described by the following stochastic differential equations:
\[
\begin{cases}
dX^{\varepsilon, \delta}_t=(AX^{\varepsilon, \delta}_t + b(X^{\varepsilon, \delta}_t, Y^{\varepsilon, \delta}_t))dt+\sqrt{\varepsilon}g(X^{\varepsilon, \delta}_t)dB^H_t, \\
dY^{\varepsilon, \delta}_t=\frac{1}{\delta}(AY^{\varepsilon, \delta}_t + F(X^{\varepsilon, \delta}_t, Y^{\varepsilon, \delta}_t))dt+\frac{1}{\sqrt{\delta}}G(X^{\varepsilon, \delta}_t, Y^{\varepsilon, \delta}_t)dW_t,
\end{cases}
\]
where \(V\) is a separable Hilbert space, \(A\) is the infinitesimal generator of an analytic semigroup, \(B^H_t\) is a cylindrical fractional Brownian motion with Hurst parameter \(H\in(\frac{1}{2}, 1)\), and \(W_t\) is a standard Brownian motion.
2. **Large Deviation Principle**:
The Large Deviation Principle focuses on the asymptotic behavior of the probability decay rate of rare events in the system when the noise intensity approaches zero. For such fast - slow systems, studying the Large Deviation Principle helps to understand the behavior of the system on a long - time scale and the probability of rare events.
3. **Insufficiency of Existing Research**:
Although there are already literatures that have studied the Large Deviation Principle for fast - slow systems driven by finite - dimensional fractional Brownian motion, at present, there is no research on the Large Deviation Principle for fast - slow systems driven by infinite - dimensional fractional Brownian motion. Therefore, this paper aims to fill this gap.
### Main Contributions
- **Innovation in Method**: Adopt the weak convergence method and deal with infinite - dimensional mixed fractional Brownian motion based on the variational representation formula.
- **Overcoming Technical Difficulties**: Through time - discretization technology and path integral estimation, overcome the problem of the lack of martingale property when the slow variable is driven by fractional Brownian motion.
- **Relaxation of Assumptions**: Relax the boundedness assumptions of the drift coefficient and the diffusion coefficient, making the results more general.
In short, this paper is committed to solving the problem of the Large Deviation Principle for fast - slow systems driven by infinite - dimensional mixed fractional Brownian motion, providing theoretical basis and technical support for further research in related fields.