The conjugacy of stochastic and random differential equations and the existence of global attractors (Q5935819)

The paper studies stochastic differential equations of the form \[ dx=f_0(x) dt+\sum_{k=1}^mf_k(x) dW_k(t)\tag{1} \] on \(\mathbb R^d\) with \(C^\infty\) vector fields~\(f_k\), \(0\leq k\leq m\), such that the~\(f_k\) are globally Lipschitz and commute for \(1\leq k\leq m\) (\text{i. e.,} \([f_k,f_n]=0\) for \(1\leq k,n\leq m\), where \([\cdot,\cdot]\) denotes the Lie bracket). Assuming the stochastic flow induced by~(1) to be forward complete, a random stationary co-ordinate transformation is constructed, such that the stochastic flow induced by~(1) is conjugate to the flow induced by a certain random differential equation. This coordinate transformation transforms attractors for~(1) to attractors for the random differential equation. This method, which had been used previously several times in the literature in a more ad hoc fashion, is then used to prove the existence of random attractors for several examples of the form~(1) (\text{viz} the Duffing-van der Pol oscillator with multiplicative noise on the position, on the velocity, and with additional additive noise, resp., the noisy damped harmonic oscillator in a double-well potential, and the Lorenz system with full and with partial multiplicative noise).