Orthogonal-Back Propagation Hybrid Learning Algorithm for Interval Type-2 Non-Singleton Type-2 Fuzzy Logic Systems

This article presents a new learning methodology based on an hybrid algorithm for interval type-2 non-singleton type-2 fuzzy logic systems (FLS) parameters estimation. Using input-output data pairs during the forward pass of the training process, the interval type-2 FLS output is calculated and the consequent parameters are estimated by orthogonal least-square (OLS) method. In the backward pass, the error propagates backward, and the antecedent parameters are estimated by back-propagation (BP) method. The proposed hybrid methodology was used to construct an interval type-2 fuzzy model capable of approximate the behavior of the steel strip temperature as it is being rolled in an industrial Hot Strip Mill (HSM) and used to predict the transfer bar surface temperature at finishing Scale Breaker (SB) entry zone. Comparative results show the advantage of the hybrid learning method (OLSBP) over that with only BP