USING GENETIC ALGORITHMS FOR AN ARTIFICIAL NEURAL-NETWORK MODEL INVERSION

Genetic algorithms (GAs) and artificial neural networks (ANNs) are tec hniques for optimization and learning, respectively, which both have b een adopted from nature. Their main advantage over traditional techniq ues is the relatively better performance when applied to complex relat ions. GAs and ANNs are both self-learning systems, i.e., they do not r equire any background knowledge from the creator. In this paper, we de scribe the performance of a GA that finds hypothetical physical struct ures of poly(ethylene terephthalate) (PET) yarns corresponding to a ce rtain combination of mechanical and shrinkage properties. This GA uses a validated ANN that has been trained for the complex relation betwee n structure and properties of PET. This technique was tested by compar ing the optimal points found by the GA with known experimental data un der a variety of multi-criteria conditions.