DCMODELSWARD 2016 Abstracts


Full Papers
Paper Nr: 1
Title:

Automatic Refactoring of Single and Multiple-view UML Models using Artificial Intelligence Algorithms

Authors:

Abdulrahman Baqais and Mohammad Alshayeb

Abstract: Refactoring tends to improve the internal structure of the software while preserving its behaviour(Fowler and Beck, 1999). This process attempts to reduce the complexity of the software and cut its maintaince cost(Mens and Tourwe, 2004) promoting its quality status(Alshayeb, 2009) . The majority of articles that discuss software refactoring are focusing on software code (Fowler and Beck, 1999, O'Keeffe and Cinnéide, 2008, Alkhalid et al., 2011b). Recently, a slight increase in the interest of refactoring at the design level emerged (Misbhauddin and Alshayeb, accepted 2013). Different methods have been applied to refactor UML diagrams namely: pattern-based (Song et al., 2002a), formal rules (Massoni et al., 2005) and graph transformation (Mens, 2006). Refactoring UML diagrams is favorable since designing activity precedes coding and as such abnormalities, ill-structure or potential bugs can be detected and corrected early (Sunyé et al., 2001). Each UML diagram has different design smells and requires different refactoring operations. From the various approaches to perform UML refactoring, there are very few targeted the advantatges, transperancy and performance that Artificial Intelligence(AI) techniques might import to the refactoring proces (Ghannem et al., 2013). Some of the issues when applying AI techinques for model refactoring (design-level refactoring) include: encoding method and selection of the right algorithm . Refactoring UML design smells for each diagram manually exhibits some drawbacks such as: it’s costly in terms of cost and time and it requires domain experts. Automating the refactoring process surely will save time and cost and will help software practitioners to improve their designs. Most of the other refactoring approaches are carried out on single instances of UML diagrams. In this research, we are extending the field by applying AI refactoring on a multiple-view UML model and comparing the results with individual UML diagrams. This will show the advantages of adopting multiple-view refactoring. The overall aim of this research is to “refactor UML models by providing the user with a set of AI techniques, that utilize software metrics and refactoring operations, to produce refactoring sequences that improve quality”.

Paper Nr: 2
Title:

Towards a Generic Multidisciplinary Models Composition Tool

Authors:

Anas Abouzahra, Ayoub Sabraoui and Karim Afdel

Abstract: The world of engineering has changed greatly in the last decades. We witnessed that softwares have become omnipresent. Various domain engineering disciplines are now strongly supported by software tools. Moreover, the increasing complexity of systems leads to develop softwares where heterogeneous engineering domains have to collaborate within the same project. The paper describes a new concept of a generic multidisciplinary models composition tool that aims to address multidisciplinary development process. We illustrate the benefits of this approach using a development use case example inspired from the actuarial science.