Genetic neuro-fuzzy architectures for advanced intelligent systems

Abstract

This paper presents a framework for developing intelligent systems based on several soft-computing techniques such as fuzzy logic, neural networks and genetic algorithm. The neural networks provide the system with a baseline structure, the fuzzy logic gives a possibility to utilize top-down knowledge from designer, and the genetic algorithm determines several system parameters with the process of bottom-up development. As a manifestation, we propose an efficient fuzzy neural system which consists of modular neural networks combined by the fuzzy integral in which genetic algorithm determines the fuzzy density values.