Parallel evolutionary algorithms for optimizing data based generated fuzzy systems
Krause, Peter, T. Slawinski, Dirk Wiesmann · 2000
. In the field of data--based fuzzy modeling, the complexity of applications and the amount of data to be processed have grown continuously. Thus, the computational effort for solving these applications has also increased drastically. In order to meet this challenge, parallel computing approaches are applied. The task here is the optimization of data--based generated fuzzy rule bases. For this kind of application the fitness evaluation of an individual is very time consuming. Here, a parallel genetic algorithm is applied to solve the optimization problem in an acceptable amount of time. Furthermore, it will be analyzed how the quality of the results changes with the use of multi--population models or neighborhood models. This will be illustrated by two example applications. 1 Introduction In recent years, many methods for the different data--mining techniques have been developed. One field of data mining is fuzzy modeling. Fuzzy modeling is quiet popular, because it is poss...