White noise eliminates instability (Q5927602): Difference between revisions

As in deterministic analysis, the concept of attractors plays a key role in the theory of random dynamical systems (RDS). For more comprehensive insight, see \textit{L. Arnold} [Random dynamical systems. Springer, New York (1998, Zbl 0906.34001)] and the author and \textit{F. Flandoli} [Attractors for random dynamical systems. Probab. Theory Relat. Fields 100, 365-393 (1994, Zbl 0819.58023)]. Now, suppose \(\varphi\) is a white noise RDS on a Riemannian manifold \(M\), \(A=A(\omega)\) is a (random) global set attractor for \(\varphi\), and there is a totally unstable invaraint measure \(\mu\) for \(\varphi\) satisfying some integrability condition like in \textit{H. Crauel} [Stochastic Process. Appl. 45, 13-28 (1993, Zbl 0778.60030)]. Then the author is able to show the existence of a compact deterministic set \(K \subset M\) such that \(K \subset A(\omega)\) (\(P\)-a.s.) as its main result. Furthermore, the author briefly explains the main notions of RDS, random attractors, Lyapunov exponents and invariant measures in this context. Finally, the example of nondegenerate stochastic differential equations (SDEs) \[ d X_t = f(X_t) dt + \sigma d W_t \] with white noise and \(\sup_{x \in \mathbb{R}^d} |f(x)|< + \infty\) is discussed to underline that \[ \forall x \in \mathbb{R}^d : P (\{ \omega \in \Omega : x \in A(\omega)\}) < 1 \] for any random attractor \(A(\omega)\) of SDEs with additive white noise. That is a deterministic point is not contained in the attractor with probability one, hence white noise perturbations in RDS may eliminate the instabilities met in the related deterministic dynamics. This could already be seen by the one-dimensional example with \(f(x) = x - x^3\) perturbed and its unstable fixed point \(0\).