Applications of unified combinatorial interaction testing

Official URL: https://risc01.sabanciuniv.edu/record=b2579387_(Table of contents)

U-CIT approach has provided a flexible and systematic method to flexibly define and compute combination interaction testing (CIT) objects generating covering arrays used in software testing. By U-CIT, software system under test and coverage criterion are expressed as a constraint solving problem, and CIT objects are computed by using appropriate constraint solvers. The convenience of defining flexibly coverage criteria brought by U-CIT has made it possible to easily define new CIT objects to test any software system. Although U-CIT objects are generated by solving constraints with constraint solvers in these studies, a higher level modelling abstraction may be required to define complex system models and coverage criteria. In this study, we present UCIT-ASP approach that we developed to generate U-CIT objects by using Answer Set Programming (ASP) which is a declarative modeling language. In addition, by using ASP modeling libraries that was developed within the scope of this study, we both generated U-CIT objects already defined in the literature (i.e. standard covering arrays, test case aware covering arrays, etc.) and defined new U-CIT objects, specifically for graph-based systems (for the testing of mobile applications, multi-threaded systems, etc.). In our case studies to experience with UCIT-ASP on the generation of well-known CIT objects, we have observed that our approach generated smaller CIT objects than specialized covering array generation methods in the literature at the cost of computation times.