A Model-Driven Approach for Simplified Cluster Based Test Suite Optimization of Industrial Systems – An Introduction to UMLTSO

Abstract

Software testing is a significant but costly activity of software development life cycle, because it accounts for more than fifty-two percent (>52%) of entire development cost. Testing requires the execution of all possible test cases in order to find defects in software. However, the selection and implementation of right test cases is always challenging for large scale industrial systems. In this context, clustering is a renowned approach for achieving optimization. However, it is difficult to optimize test cases through clustering due to its implementation complexity and time-consuming nature. Hence, a model based simple mechanism is strongly needed for optimization of generated test cases while preserving the coverage criterion. In this paper, a Unified Modeling Language profile for Test Suite Optimization (UMLTSO) is presented that models the optimization process for test case generated from java source code. Particularly, UMLTSO is capable of modeling test case generation, coverage criteria application and optimization using clustering approaches. This offers the rationale for converting the UMLTSO source code into target test cases for optimization based on different coverage criteria e.g. code coverage. The applicability of UMLTSO is validated through two industrial case studies.