Two real-time expert systems for the monitoring and maintenance of digital exchanges

The ever-increasing complexity of the elements that constitute a telecommunications network as well as the amount of knowledge and experience needed to operate and maintain these elements, demand maintenance tools that are both powerful enough to perform these tasks and easy to learn and modify. Thus, the use of a knowledge-based expert systems (KBES) methodology is required. This paper presents UFMS12 and UFMAXE, two real-time expert systems for performing automatic corrective maintenance of large digital exchanges: Ericsson's AXE and Alcatel's S12. These expert systems can repair automatically a fault dialoguing with the affected exchange or they can guide the operator in manual interventions. These developments produced a shell, called SIGMA (Sistema Generico para Mantenimiento), based on a causal model. This paper presents the functionality of UFMAXE and UFMS12, and a description of the architecture of the SIGMA shell, designed for representing the maintenance knowledge of large and complex digital systems.< >