Information State for Robust Control of Set-Valued Discrete Time Systems

We consider the problem of robust output feedback control of set-valued discrete time systems. Such systems appear frequently in applications such as process control. The aim is to motivate and obtain an efficient formulation of robust control problems and the resulting structure of the controller for such systems. This formulation, and the controller structure are obtained by taking small noise limits of the information state dynamics arising in a related risk-sensitive stochastic control problem. The necessary and sufficient conditions for solvability are also presented.