Authors:
A. ferdjoukh
;
A. Baert
;
E. Bourreau
;
A. Chateau
;
R. Coletta
and
C. Nebut
Affiliation:
LIRMM and Université Montpellier 2 and CNRS, France
Keyword(s):
Automated Model Generation, Constraint Satisfaction Problem (CSP), Object Constraint Language (OCL).
Related
Ontology
Subjects/Areas/Topics:
MetaModeling
;
Models
;
Paradigm Trends
;
Software Engineering
Abstract:
The automated generation of models that conform to a given meta-model is an important challenge in Model
Driven Engineering, as well for model transformation testing, as for designing and exploring new meta-models.
Amongst the main issues, we are mainly concerned by scalability, flexibility and a reasonable computing
time. This paper presents an approach for model generation, which relies on Constraint Programming.
After the translation of a meta-model into a CSP, our software generates models that conform to this meta-model,
using a Constraint Solver. Our model also includes the most frequent types of OCL constraints. Since
we are concerned by the relevance of the produced models, we describe a first attempt to improve them. We
outperform the existing approaches from the mentioned point of view, and propose a configurable, easy-to-use
and free-access tool, together with an on-line demonstrator.