Multiobjective dynamic optimization of semibatch copolymerization reactors

Polymerization reactor optimal design and control problems are inevitably multiobjective in nature. Copolymerization problems encompass an even larger set of objective functions than homopolymerization. This work describes in some detail a set of typical components of the vector of objective functionals for copolymerization reactor problems. A subset of these is selected for detailed examination employing multiobjective function optimization techniques. Noninferior, or Pareto, sets of optimal solutions have been determined for the dual objectives of narrowing both copolymer composition and molecular weight distributions in the styrene-acrylonitrile system. Noninferior solutions occur when the two objectives are in conflict and no objective can be improved without sacrifice of the other. We show for several sets of control variables that, in the S-AN system, one should seldom, if ever, control with respect to one objective functional exclusively. Compromise policies can be found which are close to utopian for both objectives. A better quantitative understanding of the trade-offs between objectives in copolymerization engineering has been achieved.