Model-Based Reasoning with Multiple Test Cases and its Application to Debugging

Today’s simulation-centric hardware development process requires to leverage quality test suites for fault localization rather than employing them solely for detecting malfunctioning. In this article we (1) propose an extension of the model-based debugging theory to address the treatment of test suites in a well-founded way and (2) relate this novel approach to an algorithmic technique known as filtering. For multiple test cases revealing a certain fault (3) we propose an iterative computation of diagnoses as an extension to Reiter’s diagnosis algorithm, and (4) notably report on empirical evaluations of this algorithm taking into account even dual-fault diagnosis.