Test case generation for classes in objects-oriented programming using grammatical evolution

Abstract

This paper proposes a dynamic test case generation approach for Object-Oriented Programming classes, using evolutionary search to find test cases that would satisfy a branch coverage criteria. Grammatical Evolution (GE) is used to search for a solution in accordance to user-specified grammar, thus making the algorithm more flexible than the traditional genetic programming. Rather than generating test cases directly, source code for an Intermediate Test Script (ITS) is generated from the grammar. It is then evaluated and translated into source code by ITS interpreter. Such a provision makes it easy to produce test cases that have object and literal reference, whereby improve the performance of GE. We've tested the proposed method with several java classes from open source projects and yielded high code coverage results. © 2012 Springer Science+Business Media B.V.