Abstract:We study semilinear rough stochastic partial differential equations as introduced in [Gerasimovi{č}s, Hairer; EJP 2019]. We provide $\mathcal{L}^p(\Omega)$-integrable a priori bounds for the solution and its linearization in case the equation is driven by a suitable Gaussian process. Using the Multiplicative Ergodic Theorem for Banach spaces, we can deduce the existence of a Lyapunov spectrum for the linearized equation around stationary points. The existence of local stable, unstable, and center manifolds around stationary points is also provided. In the case where all Lyapunov exponents are negative, local exponential stability can be deduced. We illustrate our findings with several examples.
What problem does this paper attempt to address?
The problems that this paper attempts to solve mainly focus on the study of solutions of semilinear rough stochastic partial differential equations (RPDEs) and their linearized \(L^p(\Omega)\)-integrable a priori bounds. Specifically:
1. **\(L^p(\Omega)\)-integrable a priori bounds**: One of the main objectives of the paper is to provide an integrable a priori bound, that is, to prove the integrability of solutions and their linearizations in the \(L^p(\Omega)\) space. This problem has not been solved in previous works, especially when the driving signal is a Gaussian process, and the known a priori bounds often do not satisfy the integrability condition. By introducing new control functions and modifying the Sewing lemma, the author has successfully solved this problem.
2. **Existence of invariant manifolds**: Using the Multiplicative Ergodic Theorem (MET) and the results of integrable a priori bounds, the paper proves the existence of local stable, unstable and central manifolds. In particular, in the case where all Lyapunov exponents are negative, local exponential stability can be deduced.
3. **Applications in stochastic dynamical systems**: The paper also explores how to apply the path - solution concept to the study of stochastic dynamical systems (RDS), especially to study the long - time behavior of these systems. This is particularly important when dealing with RPDEs driven by Gaussian processes, because the traditional stochastic integral theory is often not applicable in this case.
### Main contributions
- **Integrable a priori bounds**: Provides \(L^p(\Omega)\)-integrable a priori bounds, which is an important breakthrough in this field.
- **Invariant manifolds**: Proves the existence of local stable, unstable and central manifolds, and in particular, when all Lyapunov exponents are negative, local exponential stability can be deduced.
- **Technical methods**: Introduces new control functions and modifies the Sewing lemma. These technical methods are of great significance for extending the non - Markovian Hörmander theory to RPDEs.
### Application examples
The paper demonstrates the applications of these theoretical results through several specific examples, such as:
- **Fractional Brownian motion**: When the driving signal is fractional Brownian motion (Hurst parameter \(H\in(1/3, 1/2]\)), the results of the paper can be used to analyze the stability of solutions and the existence of invariant manifolds.
- **Stochastic dynamical systems**: By applying the path - solution concept to stochastic dynamical systems, the paper shows how to study the long - time behavior of these systems, especially in the presence of stochastic fixed points.
In conclusion, by providing integrable a priori bounds and proving the existence of invariant manifolds, this paper provides an important theoretical basis and technical tools for the study of rough stochastic partial differential equations.