Leveraging variability modeling to address metamodel revisions in Model-based Software Product Lines

Jaime Font, Lorena Arcega, Øystein Haugen, Carlos Cetina
2017 Computer languages, systems & structures  
S OFTWARE Product Lines (SPLs) exploit commonalities across a family of related products in order to increase quality and reduce time to market and costs. Most SPLs are built from a set of existing products, that needs to be re-engineered into reusable assets following feature location approaches. Traditional feature location approaches target program code, but less attention has been paid in the literature to other software artifacts such as the models. In this dissertation we present an
more » ... ch for Feature Location in Models that relies on an Evolutionary Algorithm (FLiMEA). FLiMEA capitalizes on experts domain knowledge to boost the feature location process and produce model fragments that properly capture the reusable units of the domain. The approach performs a search (guided by a fitness function) over alternative model fragment realizations of the feature being located (generated through genetic operations). As a result, variability and commonalities are formalized in the form of reusable model fragments. We have explored different genetic operations and fitness functions so the approach can be tailored to work under the different conditions present in industrial scenarios. In addition, when the features have been located and formalized as reusable assets, there is a need for evolution of those elements. In this dissertation we focus on the co-evolution of the model fragments and the language used to create them. To address this challenge we propose Variable MetaModel (VMM), an approach that relies on variability modeling ideas applied at metamodel level to enable the co-evolution. The VMM expresses each evolution of the language in terms of commonalities and variabilities, to ensure the conformance of model fragments with the new version of the language. The approaches have been validated and evaluated in our industrial partners (BSH, the biggest manufacturer of home appliances in Europe, and CAF, an international provider of railway solutions). ACKNOWLEDGEMENTS The work presented in this thesis were carried out at the
doi:10.1016/j.cl.2016.08.003 fatcat:5ahu3czexnh2pcjqctxngtenqy