Necessary and sufficient conditions for optimality in nonlinear distributed parameter systems with variable initial state (Q2638509)

We consider a nonlinear distributed control system, with time varying control constraints and an initial condition which is not determined by an a priori given function, but instead it is assumed to belong to a certain specified set (\textit{J.-L. Lions} [in: Mathematical control theory, Banach Cent. Publ. 14, 299-311 (1985; Zbl 0581.49005)] calls them ``systems with insufficient data''). The cost criterion is a general convex integral functional. Using the Dubovitskij-Milyutin formalism, we are able to obtain a necessary and sufficient condition for the existence of an optimal solution. A very comprehensive presentation of the Dubovitskij-Milyutin theory can be found in the monograph of \textit{I. V. Girsanov} [``Lectures on mathematical theory of extremum problems'' (New York 1972); for a review of the 1970 Russian original see Zbl 0214.145]. Our result extends Theorem 2.1 of Lions (loc. cit.), since we allow for nonlinear dynamics and a nonquadratic cost criterion.