A Heuristic for Near-Optimal Troubleshooting Using AO*

When troubleshooting malfunctioning technical equipment, the task is to locate faults and make repairs until the equipment functions properly again. The AO* algorithm can be used to find troubleshooting strategies that are optimal in the sense that the expected cost of repair is minimal. We have adapted the AO* algorithm for troubleshooting in the automotive domain with limited time. We propose a new heuristic based on entropy. By using this heuristic, near-optimal strategies can be found within a fixed time limit. This is shown in empirical studies on a fuel injection system of a truck. In these results, the AO* algorithm using the new heuristic, performs better than other troubleshooting algorithms.