Exponential mean square stability of stochastically forced 2-torus

From MaRDI portal





Variety in the behaviour of nonlinear dynamic systems under transition from order to chaos is connected frequently with a chain of bifurcations: a stationary regime (equilibrium point) -- periodic regime (limit cycle) -- quasiperiodic regime (torus) -- chaotic regime (strange attractor). Each such transition is accompanied by the loss of stability of a simple attractor and the birth of a new, more complicated stable attractor. Stability analysis of appropriate manifolds is the key for understanding the complex nonlinear dynamic phenomena. Cycles and stationary points are well-investigated objects of modern stability theory. At the same time, appropriate analysis of toroidal invariant manifolds is far from complete. In investigations on the stability of a stationary point the problem of Lyapunov function construction for a nonlinear system is connected traditionally with Lyapunov function construction for an appropriate linear system of the first approximation. The first approximation systems for stochastic nonlinear systems with limit cycles and their stability analyses are considered by the author [J. Appl. Math. Mech. 60, 579--590 (1996); translation from Prikl. Mat. Mekh. 60, No. 4, 582--594 (1996; Zbl 0886.93069)]. The present paper is devoted to the extension of this technique for the more complicated 2-torus manifold case. The author analyses the first approximation systems connected with the system of Itô's stochastic differential equations \[ dx = f(x)\,dt+\sum_{r=1}^m\sigma_r(x)\,dw_r(t),\tag{1} \] where \(w_r(t)\) \((r = 1,\dots,m)\) are independent standard Wiener processes, \(\sigma_r(x)\) are sufficiently smooth vector-functions of the appropriate dimension. To ensure a 2-torus \(\Gamma\) is an invariant of the stochastic system (1), it is assumed that \(\sigma_r| _\Gamma=0\). The characteristic feature of these stochastic systems is the existence of two linear independent deterministic solutions. In connection with the projector \(P\) accompanying these two solutions, a concept of exponential mean square \(P\)-stability for such linear systems is introduced. A necessary and sufficient condition of this \(P\)-stability is the existence of a solution for a matrix differential Lyapunov equation. The 2-torus stability criterion is represented in a standard form of the theorem based on the first approximation system stability. The spectral theory of positive operators is used to the analysis of the stability of toroidal motions. On the basis of a spectral approach, the parametric criterion of 2-torus stability for the case of three-dimensional space is obtained. The stochastic stability analysis is reduced to the estimation of the spectral radius of some positive operator. An example of the application of the theoretical results is considered.











This page was built for publication: Exponential mean square stability of stochastically forced 2-torus

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4810820)