\(k\)-capture in multiagent pursuit evasion, or the lion and the hyenas (Q393889)

In an \(m\)-dimensional Euclidean space the so-called \(k\)-capture problem is considered. The players' motions are described by the following discrete equations: \[ \begin{alignedat}{2} e(t+1) &= e(t)+u_{e}(t,p_1(t),\dots,p_n(t)), &&t=0,1,\dots,\\ p_j(t+1) &= p_j(t)+u_{p_j}(t,e(t),e(t+1),p_1(t),\dots,p_n(t)),\qquad &&j=1,2,\dots,n,\end{alignedat} \] where \(u_{e}\) and \(u_{p_j}\) are strategies of the evader \(e\) and the pursuer \(p_j,\) respectively. The evader is \(k\)-captured if there exists a finite time \(T\) and a subset \(C\subset \{p_1,\dots,p_n\}\) of \(k\)-pursuers such that \(\|p_j(T)-e(T)\|=0\) and \(\|p_j(T)-e(T)\|>0\) for all \(t<T\), \(p_j\in C\). In this work, necessary and sufficient conditions for \(k\)-capture are obtained.