The public defense of Eduard Paul Enoiu’s doctoral thesis in Computer Science and Engineering
Doctoral thesis and Licentiate seminars
The public defense of Eduard Paul Enoiu’s doctoral thesis in Computer Science and Engineering will take place at Mälardalen University (Västerås Campus) room Pi, at 13.15 PM on November 22, 2016.
The title of the thesis is “Automatic test generation for industrial control software”.
The faculty examiner is Professor Mats Heimdahl, University of Minnesota; and the examining committee consists of Professor Ana Rosa Cavalli, Télécom SudParis ; Professor Marjan Sirjani, MDH; Docent Christian Berger, Chalmers University of Technology.
Reserve; Professor Konstantinos Kyprianidis, MDH
The doctoral thesis has serial number 214.
As software - in different shapes, sizes and forms - is becoming an increasingly prominent part of our everyday lives, we become more and more dependent on its quality and reliability. However, software testing is more complex, time-consuming and expensive than ever before. In this context, test automation will become essential to not only meeting market demands on fast delivery but also in order to achieve high software quality. Thus far, research has provided industry with a plethora of techniques for automated software testing, but the huge amount of resources needed to automatically find suitable tests is a practical limitation in using these techniques. This limitation is of utmost importance, especially when developing industrial safety-critical software, such as the one found in power plants, airplanes, or trains.
Eduard Enoiu, a PhD Student at Mälardalen University, has devised an automatic test generation tool called CompleteTest that avoids many of these problems. The technique implemented in the tool and described in his PhD thesis, works with industrial control software and has been demonstrated to work for the detection of software faults in an industrial train control and management system produced by Bombardier Transportation. In many cases, CompleteTest can provide tests as capable of finding as many defects as those developed by an experienced human tester and could be used as an effective complementary solution to manual testing.
In addition to the tool itself, Eduard’s thesis presents an evaluation of CompleteTest, obtained through a series of studies performed in academia and industry. Results show that automatic test generation through CompleteTest is viable for use in practice; it is efficient in terms of time required to generate tests and scales well for most of the industrial programs considered. Eduard found that automatic test generation can exercise the logic of the software equally well as manual testing performed by industrial engineers but in a fraction of the time. However, Eduard also found that manually created tests are able to detect more faults of certain types than automatically generated tests. Eduard’s thesis results highlight the need for improving the goals used by automatic test generation tools by considering generating tests that are detecting specific fault types. There are still issues to resolve before CompleteTest can be applied to other type of software and more complex programs, but Eduard and his collaborators are already working on ways to overcome them. In particular, they need to understand how its usage in practice can vary depending on software complexity, human and software process factors.