Fuzzy intelligent hybrid system application to routing control in telecommunication networks

The authors propose an intelligent hybrid system architecture for real-time routing control in multipriority telecommunication networks. The intelligent hybrid system (IHS) architecture integrates the computational paradigms of an expert system and a neural network. The expert system component of IHS is a fuzzy expert system that allows for uncertainty management. The learning capability of the system is provided by the integrated neural network. The transfer of knowledge between the fuzzy expert system and the neural network is bidirectional so that the neural network is used to acquire new knowledge from the environment whereas the fuzzy expert system makes use of knowledge it had already acquired. Therefore, the fuzzy expert system relieves the neural network from learning things already known.<>