This paper presents results obtained for the control of set-valued dis
crete-time dynamical systems. Such systems model nonlinear systems sub
ject to persistent bounded noise. A robust control problem for such sy
stems is introduced, The problem is formulated as a dynamic game, wher
ein the controller plays against the set-valued system. Both necessary
and sufficient conditions in terms of (stationary) dynamic programmin
g equalities are presented. The output feedback problem is solved usin
g the concept of an information state, where a decoupling between esti
mation and control is obtained. The methods yield a conceptual approac
h for constructing controlled-invariant sets and stabilizing controlle
rs for uncertain nonlinear systems.