Conjugate duality and the control of linear discrete systems (Q2442704)

The goal of the paper is to develop a duality theory for the problem of minimizing a convex function over the solution set of a range inclusion problem determined by a multivalued operator with a convex graph, as it occurs, for example, in optimal control when the constraint set is described by linear discrete systems or differential inclusions. A dual problem is assigned to this minimization problem and, by means of conjugate duality techniques, suitable regularity conditions and a strong duality result can be provided, i.e., a result which guarantees the existence of a solution of the dual problem and the identity of the optimal objective function values of the primal-dual pair. Moreover, by use of results on convex subdifferentials, necessary and sufficient optimality conditions for the primal problem can be derived. At last, the incidence of so-called reverse strong duality is discussed: that the primal problem has a solution and the optimal values of both objective functions coincide. The obtained results are considered in the context of optimal control of linear discrete systems. Some known results are found to be particular cases of the given general results.