Dumitrescu [1-5] proposed two fuzzy training procedures (Fuzzy Perceptron and the Fuzzy Relaxation). In [3] some evolutionary algorithms for tuning these training procedures are given. The aim of this paper is to review the approach from [3] and to give some parallelized algorithms concerning this approach, using the processing capabilities of a high performance processor cluster. The paper is structured in three main sections. The first section presents the two training procedures. The second one is a short presentation of our approach on evolutionary algorithms. In the third section the results obtained by running the parallel algorithms are presented. Conclusions are drawn in the final section.
展开▼
