A hybrid diagnosis approach combining black-box and white-box reasoning

We study model-based diagnosis and propose a new approach of hybrid diagnosis combining black-box and white-box reasoning. We implemented and compared different diagnosis approaches including the standard hitting set algorithm and new approaches using answer set programming engines (DLV, Potassco) in the application of Euler/X toolkit, a logic-based toolkit for alignment of multiple biological taxonomies. Our benchmarks show that the new hybrid diagnosis approach runs about twice fast as the black-box diagnosis approach of the hitting set algorithm.