Coverage-Based Dynamic Mutant Subsumption Graph

Abstract

Dynamic mutant subsumption graph (DMSG) provides a black-box definition for redundant mutants in finite set of tests. However, the graph is proved to be complicated and unreadable. In this paper we propose coverage-based subsumption graph, a novel model that describes the structure of DMSG based on coverage. That is: 1) mutants with the same coverage are composed of a local subsumption graph; 2) subgraphs of coverage are further connected together. In comparison with DMSG, this model is believed to be more readable and thus provides a coverage-based approach to generate the mutant subsumption efficiently.