Neural Network Tools for Alarm Generation in the Telephone Network Traffic Management

Real-time network traffic management is becoming more and more crucial in Telecommunications. In this paper, we propose a modular system based on neural networks, designed to help the operator for performing diagnosis at different levels of the network. In particular, we focus on local alarm generation at the secondary transit centres of a telephone network. We finally consider some future extensions to this system for producing intelligent alarms and performing a global diagnosis.