Multiobjective dynamic optimization of discrete-time systems

A new method for multiobjective dynamic optimization of discrete time systems is proposed. As a result a stagewise algorithm for such a problem is obtained. The algorithm solves a one stage multiobjective programming subproblem with general constraints at each time step. Hence, the dimension of the decision variables in each subproblem is smaller than in the initial problem. Also, we prove that the policy obtained with the proposed method satisfies the necessary optimality conditions for that problem. Finally a case study concerning a continuous binary distillation column with some numerical results is presented.