Publicação
Multiobjective optimization of polymer extrusion: decision making and robustness
| Resumo: | A Multi-Objective Evolutionary Algorithm (MOEA) is used to optimize polymer single screw extrusion. In this approach, the MOEA is linked to a modelling routine that quantifies the objectives as a function of the decision variables (i.e., operating conditions and/or screw geometry). Due to the conflicting nature of some objectives, the optimization algorithm uses a set of possible solutions to the problem that evolves during suc-cessive generations to a set of optimal solutions denoted as Pareto set. Since practical process optimization should yield a single solution, it is convenient to implement also a Decision Making (DM) strategy. Two methodologies were followed. In one case, the solutions were selected based on the preferences of a decision maker. Alternatively, the sensitivity of the solutions to small changes in the design variables was taken into account through a robustness analysis. The analysis of various case studies and the comparison with experi-mental data validated the method and demonstrates its potential. |
|---|---|
| Autores principais: | Gaspar-Cunha, A. |
| Outros Autores: | Covas, J. A.; Denysiuk, Roman; Recio, Gustavo |
| Assunto: | Optimization Extrusion |
| Ano: | 2016 |
| País: | Portugal |
| Tipo de documento: | comunicação em conferência |
| Tipo de acesso: | acesso aberto |
| Instituição associada: | Universidade do Minho |
| Idioma: | inglês |
| Origem: | RepositóriUM - Universidade do Minho |
| Resumo: | A Multi-Objective Evolutionary Algorithm (MOEA) is used to optimize polymer single screw extrusion. In this approach, the MOEA is linked to a modelling routine that quantifies the objectives as a function of the decision variables (i.e., operating conditions and/or screw geometry). Due to the conflicting nature of some objectives, the optimization algorithm uses a set of possible solutions to the problem that evolves during suc-cessive generations to a set of optimal solutions denoted as Pareto set. Since practical process optimization should yield a single solution, it is convenient to implement also a Decision Making (DM) strategy. Two methodologies were followed. In one case, the solutions were selected based on the preferences of a decision maker. Alternatively, the sensitivity of the solutions to small changes in the design variables was taken into account through a robustness analysis. The analysis of various case studies and the comparison with experi-mental data validated the method and demonstrates its potential. |
|---|