A Rough-Neuro Fuzzy Network applied to polymer processing

There is an increasing tendency in the worldwide automotive market to consume polymeric materials, because of their processability and low cost in high volumes. This disposition gives rise to search for technological solutions in order to improve the material performance, even on the project product stage. The purpose of this paper is to predict the cycle time of an injected part according to its molding parameters using a Rough-Neuro Fuzzy Network. The methodology involves the application of Fuzzy Sets to define inference morphology in order to insert the human knowledge about polymer processing into a structured rule bases. The attributes of the molding parameters are described using membership functions and converted on Fuzzy rules. The Rough Sets Theory identified which attributes and Fuzzy relation had more influence on Artificial Neural Network (ANN) surface response. Thus, rule bases filtrate by Rough Sets were used to train a back programmed Radial Basis Function (RBF) and/or a Multilayer Perceptron (MLP) Neuro Fuzzy Network. In order to measure the performance of the proposed Rough-Neuro Fuzzy Network, the responses of the unreduced rule basis are compared with the reduced rule basis. The results show that by making use of the Rough-Neuro Fuzzy Network, it is possible to reduce the need for expertise in the construction of the Fuzzy inference mechanism.