InterCriteria Analysis by Pairs and Triples of Genetic Algorithms Application for Models Identification

摘要

In this investigation the InterCriteria Analysis (ICrA) approach is applied. The apparatuses of index matrices and intuitionistic fuzzy sets are at the core of ICrA. They are used to examine the influences of two main genetic algorithms (GA) parameters-the rates of crossover (xovr) and mutation (mutr). A series of parameter identification procedures for 5. cerevisiae and E. coli fermentation process models is fulfilled. Twenty GA with different xovr and mutr values are applied. Relations between ICrA criteria-GA parameters and outcomes, on the one hand, and fermentation process model parameters, on the other hand, are investigated. The ICrA approach is applied by pairs, as well as by triples. The obtained results are thoroughly analysed towards computation time and model accuracy and some conclusions about the derived criteria interactions are reported.