Some aspects of control systems as dynamical systems (Q686319)

The authors consider control systems of the form (*) \(\dot x=X_ 0(x)+ \sum_{i=1}^ m u_ i(t)X_ i(x)\) where \(x\in M\) (a smooth manifold), \(u=(u_ 1,\dots, u_ m)\in{\mathcal U}\) (the space of locally integrable maps \(u:\mathbb{R}\to U\), where \(U\) is a compact convex subset of \(\mathbb{R}^ n\)) and \(X_ 0,\dots,X_ m\) are smooth vector fields. They show that (*) can be associated with a dynamical system \(\Phi\) on \({\mathcal U}\times M\). This is defined as \(\Phi(t,u,x)= (\theta_ t u,\varphi(t,x,u))\) where \(\theta_ t(u(s))= u(s+t)\) and \(\varphi\) is the solution of (*) corresponding to \(u(s)\) and the initial state \(x\). The authors analyze the correspondence between the controllability properties of (*) (control sets, chain control sets, complete controllability) and certain recurrence properties of \(\Phi\) (topological mixing, topological transitivity, recurrence and chain recurrence). When \(M\) is compact, invariant measures of \(\Phi\), their ergodic decomposition and supports are also investigated.