Estimation of distribution algorithms for testing object oriented software

Abstract

Abstract-One of the main tasks software testing involves is the generation of the test cases to be used during the test. Due to its expensive cost, the automation of this task has become one of the key issues in the area. While most of the work on test data generation has concentrated on procedural software, little attention has been paid to object oriented programs, even so they are a usual practice nowadays.We present an approach based on Estimation of Distribution Algorithms (EDAs) for dealing with the test data generation of a particular type of objects, that is, containers. This is the first time that an EDA has been applied to testing object oriented software. In addition to automated test data generation, the EDA approach also offers the potential of modelling the fitness landscape defined by the testing problem and thus could provide some insight into the problem. Firstly, we show results from empirical evaluations and comment on some appealing properties of EDAs in this context. Next, a framework is discussed in order to deal with the generation of efficient tests for the container classes. Preliminary results are provided as well.

Search based software testing of object-oriented containers

“…The use of Estimation of Distribution Algorithms on this problem has been investigated with a collaboration with Sagarna [45].…”

A Univariate Marginal Approach for Pairwise Testing of Software Product Lines

“…EDAs have been adopted to solve optimization problems in software engineering. We noticed that EDAs have been utilized in optimizing test data generation and test suites generation [17], [18], [19] and fault detection [20]. EDA has also been employed to improve software reliability prediction [21].…”

On Test Data Generation of Object-Oriented Software

“…Our work [1,2,25] has focused on the testing of Container Classes (e.g., Vector, Stack and Red-Black Tree). Besides analysing how to apply different search algorithms (Random Search, Hill Climbing, Simulated Annealing, Genetic Algorithms, Memetic Algorithms and Estimation of Distribution Algorithms) to the problem and exploiting the characteristics of this type of software to help the search, we introduced more general techniques that can be applied to any OO software.…”