Optimal impulse control synthesis for a linear dynamical system (Q1326868)

This paper deals with the optimality problem of terminal impulse control of a linear dynamical system described by \[ \dot x= Ax,\quad x(0)= x_ 0\quad (x\in \mathbb{R}^ n,\;A\in \mathbb{R}^{n\times n}) \] using the control actions in terms of impulse intensity \(u_ k\) \[ u(t)= \begin{cases} 0, &\text{if}\quad t\not\in T_ u= \{t_ k, k\in K\}\\ u_ k, &\text{if} \quad t= t_ k, \quad k\in K= \{1,2,\dots, k^*\}\end{cases} \] so that at each point \(t_ k\) the function \(x(t)\) has a jump \(x(t_ k)= x(t_ k+ 0)= x(t_ k- 0)+ bu_ k\), where \(b\) is \(n\times 1\) parameter vector, and the piecewise continuous trajectory \(x(t)\), \(t\in T= [0, t^*]\) satisfies the terminal constraint \(Hx(t^*)=g\), where \(g\in \mathbb{R}^ m\), \(H\in \mathbb{R}^{m\times n}\), \(\text{rank } H= m< n\). The optimal control \(u^ 0(t)\), \(t\in T\), is determined via maximizing the criterion \[ J(u)= c' x(t^*) \] under the constraints: \(0\leq u_ k\leq u^*\), \(\sum_{k\in K} u_ k\leq \vartheta\). A theorem on the necessary conditions for this impulse-intensity optimality problem and an algorithm for real-time optimal regulator are proposed.