Oscillatory conflict-control processes (Q1316693)

The paper deals with the controlled systems \[ \dot z_ i= A_ i z_ i+ \varphi_ i (u_ i,v),\quad z_ i(0)= z_ i^ 0, \] \[ z_ i\in R^{n_ i},\quad u_ i\in U_ i,\quad v\in V,\quad i\in 1,\dots,\nu, \] where \(A_ i\) are square matrices of order \(n_ i\), \(U_ i\) and \(V\) are compact sets, \(\varphi_ i\) are continuous functions, \(u_ 1,\dots,u_ \nu\) are controls of the pursuer, \(v\) is a control of the evader. The pursuit problem is considered in the class of strategies of the form \(u_ i(t)= u_ i(v_ t(\cdot))\), where \(v_ t(\cdot)= \{v(s): s\in [0,t]\}\). It is said that pursuer's strategy ensures a capture at an instant of time \(T\) if for any measurable evader's control \(v(s)\in V\) there exists \(i\in 1,\dots,n\) such that \(z_ i(T)\in M^ i\), where \(M^ i\subset R^{n_ i}\), \(i\in 1,\dots,\nu\) is a given collection of terminal sets. Under the assumption that Pontryagin's conditions [\textit{L. S. Pontryagin}, Dokl. Akad. Nauk SSSR 174, No. 6, 1278-1288 (1967)] hold in certain intervals of time, the authors obtain sufficient conditions for the pursuit problem to be solved. Illustrating examples are presented.