A Novel Double T-wise Approach to Root Cause Analysis via Combinatorial Test Set Completion

Abstract

Combinatorial Test Design (CTD) based fault localization provides significant value in the end-to-end testing flow to identify the attribute values which are the root cause of the one or more faults. Multiple CTD model test sets need to be generated iteratively to accurately provide this fault isolation. Using the novel double t-wise set process, outlined in this paper, one can create a fault isolation complete set during the generation of a model. This double t-wise set brings fault localization into real-time execution, removing the need for a CTD model test set generator. This allows root cause analysis (RCA) and inverse CTD to exist independently after generation; providing significant benefit to hardware testing, edge computing, and CTD-based first failure data capture (FFDC) critical scenarios.