A Multiobjective Dynamic Optimization Approach for a Methyl-Methacrylate Plastic Sheet Reactor Mart  ın Rivera-Toledo,* Antonio Flores-Tlacuahuac A multiobjective dynamic optimization problem using conflicting performance objectives in polymerization systems is formulated. We use the dynamic one-dimensional mathematical model of a methyl-methacrylate cell cast reactor featuring monomer conversion and molecular weight distribution as the conflicting objectives. The aim is to compute the whole set of trade-off solutions comparing the performance of three well known procedures for addressing the solution of multi- objective optimization (MO) problems: normal boundary intersection, weighted sum, and epsi- lon-constraint. Using the air temperature profile as the manipulated variable, we demonstrate the dynamic optimal solutions obtained using the best trade-off solution from each one of the MO techniques. 1. Introduction One of the main issues addressed in the optimization of chemical processes so far has been optimization for one objective at a time. However, many practical applications involve several objectives to be considered simultaneously, for instance, the structure and manufacturing processes of large composite plastic parts, [1] design of bioprocesses, [2] biochemical reaction network, [3] for drawbead design in sheet metal forming, [4] and polymerization processes. [5,6] The appropriate objectives for a particular application are often conflicting, which means achieving the optimum for one objective requires some compromise on one or more other objectives. Multiobjective optimization (MO), partic- ularly outside engineering, refers to finding values of decision variables which correspond to and provide the optimum of more than one objective. MO can be applied to handle conflicting performance objectives. The goal is to obtain a set of equally good solutions, known as Pareto optimal solutions. In a Pareto set, no solution can be considered to be better than any other solution with respect to all objective functions. When one moves from one Pareto solution to another, at least one objective function improves while at least one other gets worse. Hence, MO involves special methods for considering more than one objective and analyzing the results obtained. The most often exploited approaches to generate this Pareto set are the weighting method and the e-constraint method, [7] although more recent approaches are available. [8] Despite surveys of MO methods are common, they are often incomplete in terms Dr. M. Rivera-Toledo, Prof. A. Flores-Tlacuahuac Departamento de Ingenier  ıa y Ciencias Qu  ımicas, Universidad Iberoamericana Prolongaci on, Paseo de la Reforma 880, M exico D.F. 01219, M exico Fax: þ52(55)59504074; E-mail: martin.rivera@uia.mx Full Paper 358 ß 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim DOI: 10.1002/mren.201300147 Macromol. React. Eng. 2014, 8, 358–373 wileyonlinelibrary.com