Using multiobjective optimization algorithms and decision making support to solve polymer extrusion problems

Abstract

Single screw extrusion is a major polymer processingoperation. Its optimization is crucial for producing goodquality products at suitable costs. This study addressesextrusion as a multiobjective optimization problem thatcan be solved using evolutionary algorithms incorporatingdecision making and robustness strategies for selectingsolutions. This approach enables focusing the search forsolutions in favored regions where the preference wasdefined either by the relative importance of the objectivesor determined considering the robustness of solutionsagainst perturbations in the design variables. The out-come of this strategy provides not only a better insightinto the problem at hand, but also facilitates the choice ofa single solution for practical implementation. The useful-ness of the approach is illustrated by several case studiesinvolving the definition of the most adequate operatingconditions, of the best screw geometry and the twotogether.