Expert maintenance systems in telecommunication networks


Our survey of some 40 network maintenance expert systems reveals theri main shortcoming, which is the difficulty to acquire troubleshooting knowledge both when initializing the expert system and after its deployment. Additionally, the state-of-the-art troubleshooting expert systems do not optimize troubleshooting cost. We present theAO * algorithm to generate a network troubleshooting expert system which minimizes the expected troubleshooting cost and learns better troubleshooting techniques during its operation.