An immune algorithm based fuzzy predictive modeling mechanism using variable length coding and multi-objective optimization allied to engineering materials processing

摘要

In this paper, a systematic multi-objective fuzzyudmodeling approach is proposed, which can be regardedudas a three-stage modeling procedure. In the first stage, anudevolutionary based clustering algorithm is developed toudextract an initial fuzzy rule base from the data. Based onudthis model, a back-propagation algorithm with momentumudterms is used to refine the initial fuzzy model. The refinedudmodel is then used to seed the initial population of anudimmune inspired multi-objective optimization algorithmudin the third stage to obtain a set of fuzzy models withudimproved transparency. To tackle the problem ofudsimultaneously optimizing the structure and parameters, audvariable length coding scheme is adopted to improve theudefficiency of the search. The proposed modeling approachudis applied to a real data set from the steel industry.udResults show that the proposed approach is capable ofudeliciting not only accurate but also transparent fuzzyudmodels.