Retour au site du LAAS-CNRS

Abstract

The main objective of this thesis is to propose a general approach for assessing the validity of a Modelling and a Simulation (M&S) used during the development of embedded systems. This approach is an effort to improve confidence in the use of a simulation whose results are often questioned without consistent justification. A contribution of this thesis is to define a framework to establish this justification. The context of the study is the set of Airbussimulation products. In relation to a Verification and Validation plan of a system, a simulation must be as close as possible to the system it represents while respecting the constraints of cost and timely availability. The simulations must be available before the systems themselves. If the level of validity is too low, the results required for the experiment can not be reached. If the level of validity is too high, time of modelling and calculation is unnecessarily spent. Considering that the validity of a simulation is never assessed in isolation but always in relation to a target user, we have defined the problem of validity as the applicability of a simulation objective of use. We have treated the problem of validity level as a hierarchy of model abstractions. We therefore propose a model for describing properties of abstractions, called the conceptual model, which is the unifying language between user and developer of the simulation. It allows, first, to speak strictly the same thing when the term "validity" is mentioned, and secondly, to assess the compatibility between an expected validity level for the experience and a validity level provided by the simulation. Then, we have established formal matching rules for mapping between an objective of use and a simulation described respectively as the conceptual model proposed. The problem of mapping is based on the principle that both are components, in the formal sense, i.e., they interact through their interfaces and only through their interfaces. We look at componentbased engineering techniques to iteratively enrich the concept of "context validity" of a model by "symbolic concepts" and cover each property of the taxonomy of abstractions. Since M&S is one process of systems engineering, we proposed a methodology to integrate our formal concepts in such a process to achieve the desired conceptual model. We illustrate this approach on an avionic communication system.

Résumé

Lobjectif de cette étude est de proposer une approche générale dévaluation de la validité dune Modélisation et Simulation (M&S) utilisée dans le cadre du développement des systèmes embarqués. Cette approche sinscrit dans une démarche visant à améliorer la confiance en lutilisation dune simulation dont les résultats sont souvent remis en cause sans justification cohérente. Le cadre dapplication de létude est lensemble des produits de simulation dAirbus. Au regard dun objectif de validation dun système, une simulation doit être la plus proche possible du système quelle représente. Dans le cycle de développement dun avion, les simulations doivent être disponibles avant les systèmes eux-mêmes. Si le niveau de validité est trop faible les résultats nécessaires à lexpérience ne peuvent être atteints. Si le niveau de validité est trop élevé, du temps de travail de modélisation et de calcul est inutilement dépensé. Nous avons assimilé le problème de niveau de validité à une hiérarchie dabstraction de modèles. Nous proposons un modèle de description des propriétés dabstractions qui permet de parler strictement des mêmes choses lorsque le terme de "validité" est évoqué et dévaluer la compatibilité entre un niveau de validité attendu par lexpérience et un niveau de validité fourni par le produit de simulation. Puis, nous avons établi des règles formelles de mise en correspondance dun objectif dutilisation et du domaine dusage dun modèle. Le problème de la mise en correspondance est fondé sur le principe quun objectif dutilisation et un domaine dusage dune simulation sont deux composants, au sens formel du terme. Nous avons adapté les techniques de lingénierie basée composants, pour enrichir, par des techniques itératives, nos deux composants. Enfin nous avons proposé une méthodologie permettant dintégrer nos concepts formels au processus dIngénierie Systèmes. Nous illustrons cette démarche sur un système de communication avionique.

Mots-Clés / Keywords

Modélisation; Simulation; Validité; Abstractions; Cadre experimental; Model; Validity; Abstraction; Component-based; Formal Matching;

Mots-Clés / Keywords

Modeling and simulation; Validation; System morphism; Conceptual model; Abstractions;

Abstract

A component-based framework for Modeling and Simulation (M&S) is proposed as a solution for handling the complex problem of analysing the validity of a M&S enabling product. Validity refers to the degree that a model and its behaviour are suitable representations of the real system and its behaviour with respect to an intended purpose of model application. A set of properties related to simulation limitations and capabilities is proposed. Structured by a formal framework, those properties allows simulation user specifying their expected degree of validity to reach their intended purpose. Then compatibility rules are specified and signature matching is used to check the compatibility between the M&S enabling product and its intended purpose.

Mots-Clés / Keywords

Validity; Abstractions; Matching; Typing;

Mots-Clés / Keywords

Abstraction level; M&S hypotheses; Valididty assessment; Matching criteria; System modeling framework;

Mots-Clés / Keywords

M&S hypotheses; Abstractions; Simplification; Validity; Experimental frame;