On the use of answer set programming for model-based diagnosis

Model-based diagnosis has been an active area of AI for several decades leading to many applications ranging from automotive to space. The underlying idea is to utilize a model of a system to localize faults in the system directly. Model-based diagnosis usually is implemented using theorem provers or constraint solvers combined with specialized diagnosis algorithms. In this paper, we contribute to research in model-based diagnosis and present a way of using answer set programming for computing diagnoses. In particular, we discuss a specific coding of diagnosis problems as answer set programs, and answer the research question whether answer set programming can be used for diagnosis in practice. For this purpose, we come up with an experimental study based on Boolean circuits comparing diagnosis using answer set programming with diagnosis based on a specialized diagnosis algorithm. Although, the specialized algorithm provide diagnoses in shorter time on average, answer set programming offers additional features making it very much attractive to be used in practice.