Integration of Functional Link Neural Networks into a Parameter Estimation Methodology

In the field of robust design, most estimation methods for output responses input factors are based on response surface methodology (RSM), which makes several assumptions regarding data. However, these may not consistently hold in real-world industrial problems. Recent studies using artificial neural networks (ANNs) indicate that input–output relationships can be effectively estimated without mentioned above. The primary objective this research is to generate a new, design dual-response method ANNs. First, second-order functional-link-NN-based approach has been proposed process mean and standard deviation (i.e., model). Second, optimal structure network defined Bayesian information criterion. Finally, functions applied compared with obtained conventional least squares (LSM)-based RSM. numerical example results imply model provide more effective solutions than LSM-based RSM, according expected quality loss criteria.