ABSTRACT
The eco-efficiency concept has been gaining relevance in the industrial reality due to a growing concern related with environmental issues. Nonetheless, the eco-efficiency is characterized by a set of metrics that often conflict, namely in the realm of manufacturing process early selection.
An approach based on metaheuristic optimization technique is proposed and applied to an injection molding case study. The results of genetic algorithms and particle swarm optimization are compared, and its integration is proposed in the developed approach model, which is applied to several common eco-efficiency ratios and a combination of them (used as cost functions).
In addition, the proposed metaheuristic approach is integrated with a process-based model that estimates the time and resources required for each alternative automatically, because in an early industrialization phase, there is no data regarding the alternative performance. The process-based model includes linear and nonlinear relations, the latter being solved using neural networks.
The results show a variety of eco-efficient solutions, where the operation of both algorithms was fruitful and their solutions complemented each other—the initial idealization of competition between algorithms switched into a cooperative action. The neural networks resulted in a reasonable accurate model.
