Controllability, integrability and ergodicity

From MaRDI portal





General continuous-time measure-preserving systems of the form \({\dot x}=f(x,u)\), \(x\in \mathbb R^n\), \(u\in \mathbb R^m\) with a drift are considered. The problem of existence of the curve \(u(t):[0,T]\to \mathbb R^m\) in \(\mathbb R^m\), such that the system moves from the initial point \(x(0)=x_0\) to the final point \(x(T)=x_T\) is described. The approach of the paper is based on using ergodic-theoretical and control-theoretical methods. The supposed program submits that ergodic properties of the drift are central in developing controllability results. The key concept is ergodic partition: partition of the phase space into subsets on which the drift is ergodic. The control leads \(x(t)\) from one ergodic subset to another. Rigorous results along these lines are proven in several interesting contexts. Some more specific results, such as an outline of the proof of sufficient and necessary conditions for controllability of \(1\)-degree and \(n\)-degree of freedom Hamiltonian systems with integrable drift, are given. The flatness method is discussed. There are comparisons with other methods for solving the problem. To some extent the method of the paper is a generalization of some of them. The author shows that the benefits of the approach to controllability based on ergodic properties of the drift is that the key steps fundamentally do not require any differential structure. The paper is rather self-contained. For the entire collection see [Zbl 1012.00049].











This page was built for publication: Controllability, integrability and ergodicity

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4452936)