A Systematic Model to Model Transformation for Knowledge-Based Planning Generation Problems

Liwen Zhang (1, 2) , Franck Fontanili (2) , Elyes Lamine (2) , Christophe Bortolaso (1) , Mustapha Derras (1) , Hervé Pingaud (3)

1 Berger-Levrault
2 CGI - Centre Génie Industriel
3 LGC - Laboratoire de Génie Chimique

The generation of an optimum planning problem and search for its solution are very complex regarding different business contexts. Generally, the problem is addressed by an optimization formulation and an optimizer is used to find a solution. This is a classical approach for the Operations Research (OR) community. However, business experts often need to express specific requirements and planning goals mathematically on a case by case basis. They also need to compute a planning result within various business constraints. In this paper, we try to support these experts during this preliminary problem design phase using a model driven engineering framework. An OR model could be generated from the knowledge included in a business conceptual model. A model to model transformation is described to support this goal. The Traveling Salesman Problem is used as a simple case study that allows explanation of our model transformation rules. We implemented our approach in the ADOxx meta-modelling Platform.

Mots clés: Operations Research; Conceptual modeling; Planning problem; ADOxx; CPLEX

Domaines: Sciences de l'ingénieur [physics]

Liwen Zhang, Franck Fontanili, Elyes Lamine, Christophe Bortolaso, Mustapha Derras, et al.. A Systematic Model to Model Transformation for Knowledge-Based Planning Generation Problems. IEA/AIE 2020 - The 33th International Conference on Industrial, Engineering & Other Applications of Applied Intelligent Systems, Sep 2020, Kitakyushu, Japan. pp.140-152, ⟨10.1007/978-3-030-55789-8_13⟩. ⟨hal-02935179⟩