Solver Representations for DOPLER Decision Models
| Typ | Masterarbeit | |
|---|---|---|
| Aushang | MA Solver Representations for DOPLER.pdf | |
| Betreuer | Wenden Sie sich bei Interesse oder Fragen bitte an: Kevin Feichtinger (E-Mail: kevin.feichtinger@kit.edu, Telefon: +49-721-608-45766) |
Motivation
Variability models are used to capture common and variable parts of a set of software(-intensive) systems. The most common approaches are of type feature model and decision model. In the last years, most research focused on feature models, so that most analysis and tooling is developed for these type of models. However, in industry, decision models are often used. Yet, existing analysis encoding are not available for decision models. Additionally, transforming decision models into feature models is not a practical solution, as there are major expressiveness differences among approaches, making the analysis results unreliable. Thus, in this thesis, existing encoding techniques for constraint optimization like CSP and ILP should be investigated and implemented for DOPLER decision models.
Tasks
Developing concepts for and compare different translations for DOPLER decision models to problem instances for solvers (ILP and CSP). Implement these translations into the existing DOPLER meta-model and evaluate the implementations with the existing translation to SMT. Show, that the encoding allows (a) identifying valid configurations of a DOPLER decision model and (b) counting the set of existing valid configurations.
Notes
This thesis can be supervised in German and/or English