Finite horizon minimax optimal control of nonlinear continuous time systems with stochastic uncertainty (Q1577519)

The paper considers stochastic uncertain systems and develops a general continuous-time stochastic framework for robustness analysis and robust control synthesis. In the authors' approach, the uncertainty in their systems has two sources. One type of uncertainty is due to an additive noise perturbation in the system. Another type of uncertainty is motivated by structured perturbations in the system dynamics. The general mathematical description of the dynamic system has the form \[ \begin{aligned} \chi(t) & = X_t(\psi(\cdot)|^t_0, m(\cdot)|^t_0)\chi_0,\quad 0\leq t\leq T,\\ \zeta(t) & = H(t,\chi(t)),\end{aligned} \] where \(\chi(\cdot)\) denots the state process, \(\psi(\cdot)\) denotes the uncertainty input process, \(\zeta(\cdot)\) denotes the uncertainty output process and \(\chi_0\) is the initial condition. Then this system is transformed into a stochastic uncertain system and the authors consider a stochastic minimax optimization problem with stochastic constraints. A general method is presented for converting problems of performance analysis or controller synthesis into unconstrained optimization problems.