Kolmogorov networks and process characteristic input-output modes decomposition

Abstract

In the past decades, representation models have been developed both via math-analytical and computational-intelligence approaches. This challenge to system sciences goes on because it involves essentially the mathematical approximation theory. Recently a comparison study via the input-output view in the time domain has been carried out. That is, an analytical decomposition representation of complex MIMO thermal processes relative to the neural-network approximation representations based on Kolmogorov's theorem. The main findings resulting out of this study are presented. These provide a novel insight as well as highlight the efficiency and robustness of fairly simple industrial digital controls, designed and implemented in the past, inherited from model approximation employed.