Search

Regarding the considerable increase of patient visits to healthcare services, especially in the Emergency Department (ED), high and rising healthcare costs continuously impact on the budget of hospitals. Optimizing the use of material and human resources is getting more and more urgent to maintain healthcare quality. In this thesis, we present appropriate solutions making for ED to improve healthcare delivery in a comprehensive vision integrating all interactions between different actors and between different patient flows. The first part concerns the analysis of ED's organizational issues in the context of French healthcare system. The second part of this thesis presents modeling and simulation of emergency treatment process, two improvement propositions are experimented using discrete events simulation tools. The third part focuses on hospital beds management; a mathematical model is developed in order to optimize the utilization of bed resource. The last part of this thesis presents firstly a detailed simulation model in which has been integrated an inpatient beds planning, simulating a scenario that the emergency department knows about inpatient beds availability in care units. Secondly, a decision support tool coupling simulation model and mathematic optimization model is discussed., L'exigence des soins à apporter aux patients accueillis aux urgences combinée au contexte économique impose aux établissements hospitaliers, surtout aux SAU, des décisions rapides ainsi que l'optimisation de l'emploi des ressources matérielles et humaines pour maintenir la fluidité des patients sans pour autant compromettre la qualité de soins. Nous proposons à travers ce mémoire de thèse des solutions appropriées aux services d'urgence permettant d'améliorer la prise en charge des urgences dans une vision globale intégrant l'ensemble des interactions entre les différents acteurs et entre les différents flux de patients. La première partie concerne l'analyse des problématiques de la filière des urgences dans un contexte socio-économique du système de santé français. La deuxième partie présente la modélisation et la simulation du processus de prise en charge des urgences, ainsi que deux propositions d'amélioration (Efficacité accrue du médecin et Parcours rapide) expérimentées. La troisième partie se focalise sur la gestion des lits d'hospitalisation en aval des urgences, un modèle mathématique est développé par la suite pour optimiser la planification des lits en tenant compte les flux des patient urgents et programmés. En dernier, nous étudions l'efficacité de la connaissance des disponibilités en lits d'aval au service d'urgence, et la planification des admissions en aval des urgences par le biais du couplage de méthodes de Simulation et de Recherche opérationnelle.

Regarding the considerable increase of patient visits to healthcare services, especially in the Emergency Department (ED), high and rising healthcare costs continuously impact on the budget of hospitals. Optimizing the use of material and human resources is getting more and more urgent to maintain healthcare quality. In this thesis, we present appropriate solutions making for ED to improve healthcare delivery in a comprehensive vision integrating all interactions between different actors and between different patient flows. The first part concerns the analysis of ED's organizational issues in the context of French healthcare system. The second part of this thesis presents modeling and simulation of emergency treatment process, two improvement propositions are experimented using discrete events simulation tools. The third part focuses on hospital beds management; a mathematical model is developed in order to optimize the utilization of bed resource. The last part of this thesis presents firstly a detailed simulation model in which has been integrated an inpatient beds planning, simulating a scenario that the emergency department knows about inpatient beds availability in care units. Secondly, a decision support tool coupling simulation model and mathematic optimization model is discussed., L'exigence des soins à apporter aux patients accueillis aux urgences combinée au contexte économique impose aux établissements hospitaliers, surtout aux SAU, des décisions rapides ainsi que l'optimisation de l'emploi des ressources matérielles et humaines pour maintenir la fluidité des patients sans pour autant compromettre la qualité de soins. Nous proposons à travers ce mémoire de thèse des solutions appropriées aux services d'urgence permettant d'améliorer la prise en charge des urgences dans une vision globale intégrant l'ensemble des interactions entre les différents acteurs et entre les différents flux de patients. La première partie concerne l'analyse des problématiques de la filière des urgences dans un contexte socio-économique du système de santé français. La deuxième partie présente la modélisation et la simulation du processus de prise en charge des urgences, ainsi que deux propositions d'amélioration (Efficacité accrue du médecin et Parcours rapide) expérimentées. La troisième partie se focalise sur la gestion des lits d'hospitalisation en aval des urgences, un modèle mathématique est développé par la suite pour optimiser la planification des lits en tenant compte les flux des patient urgents et programmés. En dernier, nous étudions l'efficacité de la connaissance des disponibilités en lits d'aval au service d'urgence, et la planification des admissions en aval des urgences par le biais du couplage de méthodes de Simulation et de Recherche opérationnelle.

This thesis deals with the analysis of the electromagnetic interactions between surface waves at HF frequencies, and a time-evolving sea surface. A simulator comprising three specific elements has been developed.The first one enables the user to model a time-varying sea surface in three dimensions, by the application of the linear wave theory. The second one simulates the interaction between the electromagnetic wave at grazing incidence and the sea surface using an exact model based on the method of moments. Lastly, a post-processing tool allows the plotting and the analysis of the resulting Doppler spectra. Validation tests are presented. Various simulations on static and dynamic surfaces (sinusoids and sea surfaces) reveal the Bragg resonance phenomenon and the Doppler effect respectively. The influence of the sea surface and the radar configuration is investigated. The simulation results show a good fit with published data and ONERA measurements. Finally, a film of pollutant on the surface is introduced in the model by the addition of a surface pressure (corresponding here to an attenuation of the heights of the waves). The effect of the presence of the film on the Doppler spectra is analyzed.; Cette thèse repose sur l'analyse des interactions électromagnétiques (EM) entre les ondes de surface, aux fréquences HF, et une surface de mer évoluant dans le temps. Un simulateur comprenant trois modules spécifiques est développé. Le premier élément permet de modéliser une surface de mer en 3D variant dans le temps, par application de la théorie linéaire des vagues. Le second fait interagir une onde EM en incidence rasante avec une surface de mer grâce à un modèle exact qui s'appuie sur la méthode des moments. Enfin, un outil de post-traitement offre la possibilité de tracer et d'analyser des spectres Doppler (SD) résultants. Une étude portant sur des surfaces statiques puis dynamiques est menée. Elle fait apparaître respectivement le phénomène de résonance de Bragg et l'effet Doppler. Des données déjà publiées ou mesurées indiquent une bonne adéquation avec les SD simulés. Finalement, l'effet sur les SD, de la présence d'un film de polluant sur la surface de mer, est analysé.

Wireless front-end receivers of last generation mobile devices operate at least two frequency translations before I/Q demodulation. Frequency translation increases the system complexity, introducing several problems associated with the mixers (dynamic range limitation, noise injection from the local oscillator, etc.). Herein, the position of the analog-to-digital in- terface in the receiver chain can play an important role. Moving the analog-to-digital converter (ADC) as near as possible to the antenna, permits to simplify the overall system design and to alleviate requirements associated with analog functions (filters, mixers). These currently re- quirements have led to a great effort in designing improved architectures as Continuous-Time Delta-Sigma ADCs. The behavioural modeling this converter, although the circuit design of the main blocks has been the subject of this thesis. The use of an advanced design methodology, al- lowing the mixed simulation at different levels of abstraction, allows to validate both the circuit design and the overall system conversion. Using a multi-feedback architecture associated with a multi-bit quantizer, the continuous-time Bandpass Delta-Sigma converter achieves a SNR of about 76 dB in a wide band of 20 MHz., La chaîne de réception des téléphones mobiles de dernière génération utilisent au moins deux étages de transposition en fréquence avant d’effectuer la démodulation en quadra- ture. La transposition en fréquence augmente la complexité du système et engendre de nom- breux problèmes tels que la limitation de l’échelle dynamique et l’introduction de bruit issu de l’oscillateur local. Il est alors nécessaire d’envisager une numérisation du signal le plus près possible de l’antenne. Cette dernière permet la conversion directe d’un signal analogique en un signal numérique à des fréquences intermédiaires. Elle simplifie ainsi la conception globale du système et limite les problèmes liés aux mélangeurs. Pour cela, des architectures moins con- ventionnelles doivent être développées, comme la conversion analogique-numérique utilisant la modulation Sigma-Delta à temps continu. La modélisation comportementale de ce convertis- seur analogique-numérique, ainsi que la conception des principaux blocs ont donc été l’objet de cette thèse. L’application d’une méthodologie de conception avancée, permettant la simulation mixte des blocs fonctionnels à différents niveaux d’abstraction, a permis de valider aussi bien la conception des circuits que le système global de conversion. En utilisant une architecture à multiples boucles de retour avec un quantificateur multi-bit, le convertisseur Sigma-Delta passe bande à temps continu atteint un rapport signal sur bruit (SNR) d’environ 76 dB dans une large bande de 20 MHz.

On-board Fault Detection, Isolation and Recovery (FDIR) systems are considered to ensure the safety and to increase the autonomy of spacecrafts. They shall be carefully designed and validated. Their implementation involves a relevant knowledge of items like functions and architectures of the system, and a fault model in relation with these items. Thus, the event-B method is well suited to correctly specify and validate on-board safety architectures. This paper focuses on the FDIR concept presentation and the use of event-B for formalising and for refining the FDIR concept. The paper is organised as follows: after a short presentation of on-board FDIR concept strongly bounded with autonomy architecture concept, we suggest activities enabling to implement FDIR concept. Then, we present the framework of formal modelling that we will use to describe our architecture and the properties related to this architecture. We illustrate our approach by modelling more specifically a safety architecture pattern that includes a primary functional component and a redundant one, under the hypothesis of no common fault. The safety property to be met is: "one single fault shall not lead to the total loss of the function". The last section of the paper deals with the objectives for the future work.

Achieved work in this thesis deals with dependability evaluation of networked controlled system (NCS). The ability of control system to offset the effects of some components' failure leads to redefine the concept of system failure. Consequently the reliability evaluation is dependent on functional parameters and becomes impossible with traditional dependability methods. This work aims at bringing a contribution relative to this aspect. To overcome these difficulties, an approach based on both modelling and simulation is proposed. We choose to work with stochastic activity network (SAN) widely used in modelling communication protocols as well as in dependability studies. First we sought to identify the incidence of two types of transient faults: loss of samples and delay within the control loop. Next we simulate the behaviour in the presence of two types of disturbances at the same time highlighting the cumulative effects. In fact the origin of the loss or delay information inside the control loop is due to the presence of the network, this aspect must be taken into account, that is why we introduce a new model representing the Controller Area Network (CAN) and injection of possible perturbations. Monte-Carlo method is used to estimate dependability parameters showing the influence of some factors such as network load for example. We have proposed a method and associated tools to approach this evaluation by simulation and thus provide assistance in designing systems to meet requirements on certain performance parameters., Ce travail s'inscrit dans le cadre de l'évaluation de la sûreté de fonctionnement des systèmes commandés en réseau (SCR). La capacité des systèmes de commandes à compenser les effets de certaines défaillances de composants amène à redéfinir le concept de défaillances du système. La conséquence est que l'évaluation de la fiabilité prévisionnelle du système est dépendante de l'évaluation fonctionnelle et devient impossible avec les méthodes traditionnelles de la sûreté de fonctionnement. Pour surmonter ces difficultés, une approche basée sur la modélisation en vue de la simulation est proposée. Nous avons choisi les Réseaux d'activités stochastiques (SAN) largement connus dans la modélisation des protocoles de communication ainsi que dans les études de la sûreté de fonctionnement. Dans un premier temps, nous avons cherché à identifier l'incidence de deux types de défaillances fugitives : la perte d'un échantillon d'une part et le retard d'un échantillon dans la boucle de régulation d'autre part. Après, nous simulons le comportement en présence des deux types de perturbations simultanément, mettant en évidence des effets cumulatifs. Si on tient compte maintenant du fait que l'origine des pertes ou retards est due à la présence du réseau, il faut l'introduire dans le modèle. On introduit alors dans le modèle global du système la représentation SAN d'un réseau CAN et l'injection des défaillances dans celui-ci. La méthode de Monte Carlo est utilisée pour estimer les indicateurs de sûreté de fonctionnement et on montre l'influence de certains facteurs comme la charge du réseau par exemple. Nous avons proposé une méthode et les outils associés pour approcher cette évaluation par simulation et ainsi apporter une aide à la conception des systèmes satisfaisant à des exigences critiques sur certains paramètres de performance.

Achieved work in this thesis deals with dependability evaluation of networked controlled system (NCS). The ability of control system to offset the effects of some components failure leads to redefine the concept of system failure. Consequently the reliability evaluation is dependent on functional parameters and becomes impossible with traditional dependability methods. This work aims at bringing a contribution relative to this aspect. To overcome these difficulties, an approach based on both modelling and simulation is proposed. We choose to work with stochastic activity network (SAN) widely used in modelling communication protocols as well as in dependability studies. First we sought to identify the incidence of two types of transient faults: loss of samples and delay within the control loop. Next we simulate the behaviour in the presence of two types of disturbances at the same time highlighting the cumulative effects. In fact the origin of the loss or delay information inside the control loop is due to the presence of the network, this aspect must be taken into account, that is why we introduce a new model representing the Controller Area Network (CAN) and injection of possible perturbations. Monte-Carlo method is used to estimate dependability parameters showing the influence of some factors such as network load for example. We have proposed a method and associated tools to approach this evaluation by simulation and thus provide assistance in designing systems to meet requirements on certain performance parameters, Ce travail s'inscrit dans le cadre de l'évaluation de la sûreté de fonctionnement des systèmes commandés en réseau (SCR). La capacité des systèmes de commandes à compenser les effets de certaines défaillances de composants amène à redéfinir le concept de défaillances du système. La conséquence est que l'évaluation de la fiabilité prévisionnelle du système est dépendante de l'évaluation fonctionnelle et devient impossible avec les méthodes traditionnelles de la sûreté de fonctionnement. Pour surmonter ces difficultés, une approche basée sur la modélisation en vue de la simulation est proposée. Nous avons choisi les Réseaux d'activités stochastiques (SAN) largement connus dans la modélisation des protocoles de communication ainsi que dans les études de la sûreté de fonctionnement. Dans un premier temps, nous avons cherché à identifier l'incidence de deux types de défaillances fugitives : la perte d'un échantillon d'une part et le retard d'un échantillon dans la boucle de régulation d'autre part. Après, nous simulons le comportement en présence des deux types de perturbations simultanément, mettant en évidence des effets cumulatifs. Si on tient compte maintenant du fait que l'origine des pertes ou retards est due à la présence du réseau, il faut l'introduire dans le modèle. On introduit alors dans le modèle global du système la représentation SAN d'un réseau CAN et l'injection des défaillances dans celui-ci. La méthode de Monte Carlo est utilisée pour estimer les indicateurs de sûreté de fonctionnement et on montre l'influence de certains facteurs comme la charge du réseau par exemple. Nous avons proposé une méthode et les outils associés pour approcher cette évaluation par simulation et ainsi apporter une aide à la conception des systèmes satisfaisant à des exigences critiques sur certains paramètres de performance

International audience; Textile composites made of woven fabrics have demonstrated excellent mechanical properties for the production of high specific-strength products. Research efforts in the woven fabric sheet forming are currently at a point where benchmarking will lead to major advances in understanding both the strengths and the limitations of existing experimental and modeling approaches. Test results can provide valuable information for the material characterization and forming process design of woven composites if researchers know how to interpret the results obtained from varying test methods appropriately. An international group of academic and industry researchers has gathered to design and conduct benchmarking tests of interest to the composite sheet forming community. Shear deformation is the dominative deformation mode for woven fabrics in forming; therefore, trellis-frame (picture-frame) and bias-extension tests for both balanced and unbalanced fabrics have been conducted and compared through this collaborative effort. Tests were conducted by seven international research institutions on three identical woven fabrics. Both the variations in the setup of each research laboratory and the normalization methods used to compare the test results are presented and discussed. With an understanding of the effects of testing variations on the results and the normalization methods, numerical modeling efforts can commence and new testing methods can be developed to advance the field.

Le thème de cette étude est la modélisation du séchage d'un milieu poreux déformable partiellement saturé (solide, liquide et gaz). Cette modélisation tient compte de la nature du produit ainsi que des conditions initiales et de sa forme. En fait, au cours de leur séchage, les milieux poreux déformables subissent des contraintes liées au retrait volumique. L'objectif est de prévoir ces contraintes afin de contrôler la déformation du produit. La modélisation du séchage des milieux poreux saturés étant maîtrisée, de même pour les milieux non saturés et non déformables. La problématique se pose au niveau de la transition entre milieu saturé et milieu non saturé où il n'existe pas actuellement de modélisation physique continue. Dans ce travail, on propose un modèle décrivant les transferts de chaleur, de masse et de quantité de mouvement appliqué au séchage d'un milieu non saturé et déformable. Le gradient de pression est le terme moteur du transport de l'eau dans le milieu au travers de la loi de Darcy. La particularité du modèle est qu'il tient compte du fort couplage entre transport et comportement rhéologique du matériau en utilisant la notion de contraintes effectives. Les variables de couplage sont la vitesse de déformation du solide et la pression intrinsèque de la phase liquide. Le modèle est validé pour une argile à différentes conditions de séchage convectif. Les simulations montrent la faisabilité du modèle décrivant le séchage d'un milieu partiellement saturé et l'étude de sensibilité montre la forte influence de la perméabilité du matériau et de la pression capillaire d'une part, et des propriétés rhéologiques d'autre part.

This paper presents some experiments and some results to enforce real time distributed simulations in accordance with the High Level Architecture (HLA). Simulations were run by using CERTI, an open source middleware, as the Run Time Infrastructure (RTI). Models were distributed over computers under various available versions of the 2.6 Linux kernel. Studies and experiments relied on a real case study. The chosen case study was the simulation of an "in formation" flight of observation satellites. This case study brings up some real applicative needs in real time distributed simulations and real configurations of simulators and models. Two simulations of "in formation" flight of satellites were studied. The study consisted in modeling the behaviour of the simulators and in running these models by using various kernel or middleware operating mechanisms and services. Time measurements were performed at each test giving some results on the ability of the simulation to meet its real time requirements.

Improving processor capabilities by increasing the number of cores produces sharing effects over computer components among core requests. Furthermore, parallel applications from scientific fields attempt tofully exploit the resources of parallel architectures. Therefore, understanding resource-sharing phenomenon becomes one major issue for actual parallel architectures.These new behaviours of resource sharing are hard to analyse and to predict. In this work, we study the problem of network sharing. Clusters use dedicated high performance networks such as Gigabit Ethernet, Myrinet or Quadrics. Simultaneous executions of application tasks create concurrent access over the network. Their effects lead to performance leaks because of bandwidth network segmentation among communications.In this context, we present a deep analysis of concurrent behaviours for Quadrics, Myrinet and Gigabit Ethernet. This analysis aims to the definition of predictive models based on the concept of bandwidth sharing. In addition, we integrate such models in a simulation process for predicting concurrent communication impacts resulting of scientific application runs. Predicting concurrent behaviour leads, in scientific and technologic areas, to a better knowledge of application needs and, in industrial context, to advance cluster solutions adjusted to customer requirements.; La croissance des capacités de calcul des processeurs se poursuit, non plus par l'augmentation des fréquences d'horloge, mais par la multiplication d'unités de traitement (coeur) au sein des chips. Cette augmentation du nombre de coeurs par processeur induit un partage des autres omposants de la machine entre les différentes requêtes. En outre, les applications parallèles provenant du calcul scientifique s'efforcent d'exploiter au maximum les ressources d'une architecture parallèle. La compréhension du partage de ressource devient un des défis majeurs pour utiliser efficacement ces architectures.Ces nouveaux comportements de partage de ressources, ainsi produits, sont difficiles à interpréter et à prédire. Dans cette thèse, nous avons étudié le problème du partage du réseau. Les grappes de calcul utilisent des réseaux dédiés tels que Gigabit Ethernet, Myrinet ou Quadrics. L'exécution simultanée des tâches d'une application entraîne des accès concurrents sur la ressource réseau. Leurs effets conduisent à une perte de performance qui découle du partage de la bande passante réseau entre communications.Suivant ce contexte, nous présentons une analyse fine des comportements concurrents sur les architectures : Quadrics, Myrinet et Gigabit Ethernet. Cette analyse conduit à la définition de modèles prédictifs basés sur la notion de partage de la bande passante. En outre, nous montons que l'intégration de ces modèles dans une simulation permet de prédire les impacts dûs à la concurrence entre communications MPI résultantes de l'exécution d'applications scientifiques. La prédiction des comportements concurrents donne lieu, au niveau scientifique et technique, à une meilleure connaissance des besoins des applications et, au niveau industriel, à la proposition de solutions de grappes adaptées aux besoins de leurs clientèles.

Improving processor capabilities by increasing the number of cores produces sharing effects over computer components among core requests. Furthermore, parallel applications from scientific fields attempt tofully exploit the resources of parallel architectures. Therefore, understanding resource-sharing phenomenon becomes one major issue for actual parallel architectures.These new behaviours of resource sharing are hard to analyse and to predict. In this work, we study the problem of network sharing. Clusters use dedicated high performance networks such as Gigabit Ethernet, Myrinet or Quadrics. Simultaneous executions of application tasks create concurrent access over the network. Their effects lead to performance leaks because of bandwidth network segmentation among communications.In this context, we present a deep analysis of concurrent behaviours for Quadrics, Myrinet and Gigabit Ethernet. This analysis aims to the definition of predictive models based on the concept of bandwidth sharing. In addition, we integrate such models in a simulation process for predicting concurrent communication impacts resulting of scientific application runs. Predicting concurrent behaviour leads, in scientific and technologic areas, to a better knowledge of application needs and, in industrial context, to advance cluster solutions adjusted to customer requirements.; La croissance des capacités de calcul des processeurs se poursuit, non plus par l'augmentation des fréquences d'horloge, mais par la multiplication d'unités de traitement (coeur) au sein des chips. Cette augmentation du nombre de coeurs par processeur induit un partage des autres omposants de la machine entre les différentes requêtes. En outre, les applications parallèles provenant du calcul scientifique s'efforcent d'exploiter au maximum les ressources d'une architecture parallèle. La compréhension du partage de ressource devient un des défis majeurs pour utiliser efficacement ces architectures.Ces nouveaux comportements de partage de ressources, ainsi produits, sont difficiles à interpréter et à prédire. Dans cette thèse, nous avons étudié le problème du partage du réseau. Les grappes de calcul utilisent des réseaux dédiés tels que Gigabit Ethernet, Myrinet ou Quadrics. L'exécution simultanée des tâches d'une application entraîne des accès concurrents sur la ressource réseau. Leurs effets conduisent à une perte de performance qui découle du partage de la bande passante réseau entre communications.Suivant ce contexte, nous présentons une analyse fine des comportements concurrents sur les architectures : Quadrics, Myrinet et Gigabit Ethernet. Cette analyse conduit à la définition de modèles prédictifs basés sur la notion de partage de la bande passante. En outre, nous montons que l'intégration de ces modèles dans une simulation permet de prédire les impacts dûs à la concurrence entre communications MPI résultantes de l'exécution d'applications scientifiques. La prédiction des comportements concurrents donne lieu, au niveau scientifique et technique, à une meilleure connaissance des besoins des applications et, au niveau industriel, à la proposition de solutions de grappes adaptées aux besoins de leurs clientèles.

Le travail présenté dans ce papier a pour objectif de proposer une méthodologie de modélisation permettant de combiner au sein d’un même modèle différents processus (sollicitation, défaillance, réparation, approvisionnement, etc.) concernant le système principal et son système de soutien. L’approche linéaire classiquement suivie dans la conduite de cet exercice met en évidence une difficulté de modélisation certaine due à la complexité et au nombre des aspects mis en jeu. Bien que l'approche apparaisse probante sur le plan des résultats, l’absence de structuration réduit la confiance susceptible d’être accordée aux modèles et implique en conséquence une attention accrue de la part du modélisateur. Pour résoudre ce problème, nous proposons une démarche structurée conduisant à la réalisation de Réseaux de Petri capables de prendre en compte à la fois l’aspect physique des systèmes étudiés mais aussi des notions plus immatérielles liées aux stratégies mises en place. La finalité de cette approche est l’évaluation globale de la performance technico-économique d’un système dans sa phase d’exploitation pour l’identification des meilleures solutions sur le plan de l’architecture des systèmes mais aussi de l’organisation mise en place pour les faire fonctionner. La démarche de structuration de la modélisation par Réseaux de Petri est d’abord présentée avant d'être appliquée au cas d’un système industriel.

This paper presents a study of some reliability mechanisms which may be put at work in the context of Satellite Digital Multimedia Broadcasting (SDMB) to mobile devices such as handheld phones. These mechanisms include error correcting codes, interleaving at the physical layer, erasure codes at intermediate layers and error concealment on the video decoder. The evaluation is made on a realistic satellite channel and takes into account practical constraints such as the maximum zapping time and the user mobility at several speeds. The evaluation is done by simulating different scenarios with complete protocol stacks. The simulations indicate that, under the assumptions taken here, the scenario using highly compressed video protected by erasure codes at intermediate layers seems to be the best solution on this kind of channel

1. Introduction Nacre (the pearly internal layer of seashells) is a natural nanocomposite currently studied for the design of new organic/inorganic hybrid materials by mimicking biomineralization processes. It is a bioceramic formed at ambient temperature and pressure [1] which displays an exceptional high strength, stiffness and toughness [2] to weight ratio, as well as a natural biocompatibility with human bones [3].

On-board FDIR (Fault Detection, Isolation and Recovery) is contemplated for autonomous satellite formations. Several FDIR strategies have been specified using the Petri net - based software ProCoSA (for the dynamic aspects) on the one hand, and the set theory - based Z specification language (for the static aspects) on the other hand. ProCoSA enables to specify the different state changes triggered by the different events within the formation; Z enables to describe the relations and constraints(invariants) between the state variables. The paper focuses on a global specification including both the dynamic and static aspects, through a formal link between ProCoSA and Z. The link is implemented and allows some properties of the strategies to be checked.

PrODHyS est un environnement de simulation dynamique hybride dont la particularité est de fournir des composants orientés objets généraux et réutilisables pour la modélisation d'appareils et d'opérations présents dans les procédés industriels. Les grands principes sur lesquels est fondé PrODHyS sont d'abord exposés en développant plus particulièrement les aspects relatifs au niveau commande. Ces concepts sont ensuite illustrés à travers la simulation d'un exemple académique de procédé batch permettant d'illustrer les potentialités du simulateur. Enfin, la dernière partie aborde brièvement des travaux en cours montrant l'exploitation de ce simulateur pour la conduite de ce type de procédés (ordonnancement et supervision).

Les travaux présentés dans ce papier reprennent l'ensemble des aspects abordés lors de l'automatisation d'un système qui se veut être autonome : la modélisation, la linéarisation, la synthèse de loi de commande, la simulation, le pilotage, le guidage et l'optimisation. Le support expérimental est un micro-drone dont le domaine de vol est élargi(de la capacité de vol stationnaire "comme un hélicoptère" au vol d’avancement rapide "comme un avion"). Les points abordés ici sont plus particulièrement, la modélisation, la linéarisation, le pilote pour le vol stationnaire et le pilote longitudinal pour les transitions de phases de vol autonomes effectuées par séquencement de gains. Nous indiquerons également la structure du simulateur complet non linéaire qui permet de tester les lois réalisées avant de les embarquer. A la fin de l'article, les perspectives et la suite des travaux seront présentées.

Classical SAR Imagery algorithms or SAR processors are all based on the isotropic point model. When detecting Man Made Targets (MMT), this assumption shows its limitation because of the directive behavior of these kind of targets: this model does not take into account their physical properties. The basic idea of this thesis is that we shall be able to develop more efficient SAR processors for the detection of MMT, provided that we adopt a more suited description for these targets. A natural way to model these MMT is to consider them as sets of canonical elements with unknown orientation. If the signal backscattered by the canonical element, whatever its orientation, belongs to a low dimensional subspace, it is then possible to develop a SAR processor based on subspace detectors matched to the canonical element. This processor is called Subspace Signal Detector SAR (SSDSAR) algorithm. When the MMT is embedded in a medium, where specific scatterers create interferences, one can model the electromagnetic response of these scatterers to clear them from the SAR image. Once again, if the set of signals backscattered by the element used to model the scatterer, whatever its orientation, belongs to a low dimensional subspace, a SAR processor based on subspace detectors can be developed. This processor is called Signal or Interference Subspace Detector SAR (SISDSAR) algorithm. Different ways to implement SSDSAR and SISDSAR are presented. Theoretical performances of the two new algorithms are studied in details. Finally, we apply these two processors to simulated and real data., Les algorithmes classiques d'imagerie SAR sont tous basés sur le modèle du point isotrope. Lors de la détection de cibles manufacturées (Man-Made Targets ou MMT), cette hypothèse montre ses limites en raison de la diffusion directive de ce genre de cibles: ce modèle ne tient pas compte de leurs propriétés physiques. L'idée de base de cette thèse est que nous serons en mesure de développer des processeurs SAR plus efficaces pour la détection du MMT, à condition que nous adoptions une description plus adaptée à ces objectifs. Une façon naturelle de modéliser ces MMT est de les considérer comme des ensembles d'éléments canoniques avec orientation inconnue. Si le signal rétrodiffusé par l'élément canonique, quelle que soit son orientation, appartient à un sous-espace de faible dimension, il est alors possible d'élaborer un processeur SAR sur la base de détecteurs de sous-espaces adaptés à l'élément canonique. Cet algorithme est appelé processeur SAR basé sur un Détecteur de Sous Espaces Signaux (SSDSAR). Lorsque le MMT est intégré dans un milieu où les diffuseurs spécifiques créent des interférences, on peut modéliser la réponse électromagnétique de ces diffuseurs pour les faire disparaître de l'image SAR. Encore une fois, si l'ensemble des signaux rétrodiffusés par l'élément utilisé pour modéliser le diffuseur, quelle que soit son orientation, appartient à un sous-espace de dimension faible, un processeur SAR basé sur des détecteurs subspatiales peut être développé. Ce processeur est appelé processeur SAR basé sur des Détecteurs de Sous Espaces Signaux ou Interférences (SISDSAR). Différentes façons de mettre en œuvre SSDSAR et SISDSAR sont présentées. Les performances théoriques des deux nouveaux algorithmes sont étudiées en détail. Enfin, nous appliquons ces deux processeurs à des données réelles et simulées.

International audience; Environmental issues, such as product recyclability, are becoming a crucial social concern for manufacturers. Often formulated in natural language, the norms and standards that govern these issues can be difficult to link to the product definition. It becomes necessary to provide the designer with appropriate support tools allowing, for instance, for the compliance of products with environmental criteria to be checked. In the current paper, we show how normative knowledge, coming from textual sources (eco-labels), can be expressed through constraints, allowing checking in a semi-automated process the recyclability of a product.

International audience; Genomics experiments are widely acknowledged to produce a huge amount of data to be analysed. The challenge is to extract meaningful biological context for proteins or genes which is currently difficult because of the lack of an integrative workflow that hinders the efficiency and the robustness of data mining performed by biologists working on ruminants. Thus, we designed ProteINSIDE, a free web service (www.proteinside.org) that (I) provides an overview of the biological information stored in public databases or provided by annotations according to the Gene Ontology, (II) predicts proteins that are secreted to search for proteins that mediate signalisation between cells or tissues, and (III) analyses protein-protein interactions to identify proteins contributing to a process or to visualize functional pathways. Using lists of proteins or genes as a unique input, ProteINSIDE is an original all-in-one tool that merges data from these searches to present a fast overview and integrative analysis of genomic and proteomic data from Bovine, Ovine, Caprine, Human, Rat, and Murine species. ProteINSIDE was bench tested with 1000 proteins identifiers from each species by comparison with DAVID, BioMyn, AgBase, PrediSi, and Phobius. Compared to DAVID or BioMyn, identifications and annotations provided by ProteINSIDE were similar from monogastric proteins but more numerous and relevant for ruminants proteins. ProteINSIDE, thanks to SignalP, listed less proteins potentially secreted with a signal peptide than PrediSi and Phobius, in agreement with the low false positive rate of SignalP. In addition ProteINSIDE is the only resource that predicts proteins secreted by cellular processes that do not involve a signal peptide. Lastly, we reported the usefulness of ProteINSIDE to bring new biological hypotheses of research from proteomics data: the biological meaning of the uptake of adiponectin by the foetal muscle and a role for autophagy during ontogenesis of adipose and muscle tissues.

cote interne IRCAM: Helie06e; None / None; National audience; Fractional differential systems are infinite-dimensional systems which are difficult to study and simulate: they can be represented with poles and cuts. This representation applies to a wider class of irrational transfer functions, and is most useful for signal processing purposes, such as frequency-domain and time-domain simulations: the approximations in low dimension which give the most striking numerical results are obtained through an optimization procedure, the parameters of which are meaningful from a signal point of view. Ten such systems of increasing complexity are thoroughly investigated.

International audience; The increasing complexity of enterprise systems requires a more advanced analysis of the representation of services expected than is currently possible. Consequently, the specification stage, which could be facilitated by formal verification, becomes very important to the system life-cycle. This paper presents a formal modelling approach, which may be used in order to better represent the reality of the system and to verify the awaited or existing system's properties, taking into account the environmental characteristics. For that, we firstly propose a formalization process based upon properties specification, and secondly we use Conceptual Graphs operations to develop reasoning mechanisms of verifying requirements statements. The graphic visualization of these reasoning enables us to correctly capture the system specifications by making it easier to determine if desired properties hold. It is applied to the field of Enterprise modelling.

Structures of power converters constitute very complex systems, the behaviour of which introduces several phenomenons (electric, thermal and mechanic). An accurate modelling of their behaviour, thus, requires a multi-physics model which makes it possible to take into account the physical phenomenon in consideration. To achieve this, a FEM code (e. g. MulPhyDo) adapted to power converters and dedicated to simulation of structure in a multi-physics context has been built in the laboratory. The aim of the present study is to include in the modelling the existence of randomness and uncertainties, particularly of interest in the case of power converters. In order to achieve this, the multi-physics FEM code is coupled with a program dedicated to structural reliability analysis, in an Oriented Object Programming context. In this way, it is possible to compute the failure reliability of the structure under consideration and the sensitivity of the reliability index to each random variable. An application example of a microelectronic package is presented.

cote interne IRCAM: Helie06d; None / None; International audience; Acoustic waves travelling in axisymmetric pipes with visco-thermal losses at the wall obey a Webster–Lokshin model. Their simulation may be achieved by concatenating scattering matrices of elementary transfer functions associated with nearly constant parameters (e.g. curvature). These functions are computed analytically and involve diffusive pseudo-differential operators, for which we have representation formula and input-output realizations, yielding direct numerical approximations of finite order. The method is based on some involved complex analysis.

L'importance croissante accordée à la simulation numérique dans l'industrie aéronautique souléve le porblème de la qualité des modèles utilisés. Dans le domaine de la simulation de la dynamique du vol, des protocoles d'essais sont conçus et réalisés sur avion réel afin de recaler les valeurs prévisionnelles des coefficients aérodynaliques du modèle. dans le contexte d'une amélioration de la précision des paramètres estimés et de la diminution des coûts des essais en vol, cet article présente une nouvelle méthodologie d'optimisation fondé sur la technique des algorithmes génétiques (AG) pour la conception de protocoles d'essais optimaux.

This paper considers the problem of modeling and analyzing the reliability of a system or a component (system) where the state of the system and the state of process variables influences each other in addition to an exogenous perturbation influence: this is the dynamic reliability. We consider discrete time case, that is the state of the system as well as the state of process variables are observed or measured at discrete time instants. A mathematical tool that shows interesting properties for modeling and analyzing this problem is the so called Dynamic Bayesian Networks (DBN) that permit graphical representation of stochastic processes. Furthermore their learning and inference capabilities can be exploited to take into account experimental data or expert’s knowledge. We will show that a complex interaction between system and process on one hand and between system, process and exogenous perturbation on the other hand can simply be represented graphically by a dynamic Bayesian network. With their extended tool, known as influence diagrams (ID) that integrate actions or decisions possibilities, one can analyze and optimize a maintenance policy and/or make reactive decision during an accident by simulating different scenarios of its evolution for instance.

This paper assesses the way error control is managed jointly by Forward Error Codes (FEC) and Cyclic Redundancy Checks (CRC) in the lower layers of today’s Digital Video Broadcasting (DVB) satellites. Mathematical and simulation results clearly show that the outer block codes of the coding schemes used in DVB-S and DVB-S2 (Reed-Solomon and Bose-Chaudhuri-Hocquenghem, respectively) can provide very accurate error-detection information to the receiver in addition to their basic correction task at virtually no cost, making an uncorrected error after decoding an extremely improbable event. For this reason, the workload of CRCs can be ensured safely by the FEC subsystem if a dedicated function allowing the physical layer to share its decoding information with the adaptation layer is set. This particular crosslayer mechanism would allow freeing up the bandwidth currently used by CRCs -which adds up to 10 % for more than 35 % of the total number of IP packets- and pave the way for an enhanced transport of IP over DVB-S2.

The growing development of numerical simulation in aviation industry raises the issue of the quality of the models implemented. Flight test protocols are designed and flown to adjust the coefficients of the predicted aerodynamic model to the real aircraft in order to improve the flight dynamics model accuracy. The aircraft model becomes of major importance in every new aircraft program. Consequently, the constraint on its reliability increases. This paper describes a new optimization methodology using Genetic Algorithms (GA) for the design of optimal flight test protocols and provides a comparison with alternative method such as Dynamic Programming (DP).

The take-off is a preponderant time that influences performance. This work focuses on the indirect measurement of external forces applied on subjects thanks to motion capture. This measurement requires the use of body segments parameters that were obtained with the averaged data of different subjects. This work proposes a method to customise body segments parameters. External forces are evaluated and then, modelled with cubic splines capturing the shapes of the corresponding curves. The resulting control points enhance the comparison between the take-offs. Both methods are used to simulate new take-offs. This simulation is obtained by optimisation of the control points to verify new constraints and new performances. As a perspective, this work could lead to new approaches to optimise performance in take-off., L'impulsion est une phase déterminante de la performance sportive. Ce travail s'intéresse ainsi à la mesure indirecte par l'acquisition de mouvements, des forces mises en jeu lors de la phase de contact avec le sol. Cette mesure fait généralement intervenir les tables anthropométriques basées sur des moyennes. Ce travail cherche à individualiser les paramètres segmentaires plutôt que d'utiliser ces tables. La résultante des forces ainsi déterminée est modélisée par des splines cubiques permettant la comparaison de son évolution. Pour cela, on utilise des paramètres qui prennent en compte la forme de la courbe. Ces deux techniques sont utilisées pour permettre la simulation de nouvelles impulsions. Cette simulation repose sur l'adaptation de nouveaux mouvements à partir d'impulsions déjà acquises. Les points caractéristiques qui modélisent l'impulsion sont alors optimisés pour adapter la performance à de nouvelles contraintes.

This work concerns the development of a virtual prototyping platform dedicated to the simulation of problems in large thermomechanical transformations. In such a situation, the material is subjected to large deformations, high strain rates and a high amont of temperature mainly resulting from plastic deformation. This category of sollicitations, which might be very dynamic and highly non-linear, is usually encountered in machining or impact and crash numerical simulations. Two main research ways are explored in this work: the first one concerns the experimental aspect while the second one focusses on the development of the FEM dynamic code called DynELA. Concerning the experimental approach, the main objective is to develop a set of tools (experimental set and usefull related softwares) in order to identify the constitutive parameters of various materials subjected to high strain and high strain rates. Concerning the numerical aspect, the objective is to develop an explicit large deformation FEM code using an Object-Oriented Programming aproach in C++. This work has been achieved in the Laboratoire Génie de Production (LGP) of the Ecole Nationale d'Ingénieurs de Tarbes (ENIT)., Ce travail concerne le développement d'une plateforme de prototypage virtuel dédiée à la simulation numérique des problèmes en grandes transformations thermomécaniques rapides. Dans ces situations, le matériau est soumis à de grandes déformations, de forts taux de déformations et une forte élévation de température résultant principalement de la déformation plastique. Cette sorte de sollicitations, pouvant être très rapides et fortement non linéaire, est généralement rencontrée lors de simulations numériques de problèmes d'usinage ou d'impact et de crash. Deux voies de recherche sont explorées dans ce travail: la première concerne l'aspect expérimental alors que la seconde se focalise sur le développement du code de calcul éléments finis dynamique DynELA. Concernant l'approche expérimentale, l'objectif principal est de développer un ensemble d'outils (dispositif expérimental et logiciels utilitaires dédiés) dans le but d'identifier les paramètres constitutifs pour divers matériaux soumis à des grandes déformations et de forts taux de déformation. Concernant l'aspect numérique, l'objectif est de développer un code de calcul par éléments finis explicite basé sur une approche de Programmation Orientée-Objets en C++. Ce travail a été réalisé au sein du Laboratoire Génie de Production (LGP) de l'Ecole Nationale d'Ingénieurs de Tarbes (ENIT).

This work concerns the development of a virtual prototyping tool for large scale electro-thermo-mechanical simulation of power converters used in railway transport including multi-domain and multiple time-steps aspects. For this purpose, Domain Decomposition Method (DDM) is used to give on one hand the ability to treat large scale problems and on the other hand, for transient analysis, the ability to use different time-steps in different parts of the numerical model. An Object-Oriented programming in C++ is used for the development of the application, allowing integration in a single interface, high evolutivity and code maintenance. Finally, parallelization is introduced to improve the performances of the code and obtain computing times compatible with industrial developments. In a first part, we present the formulation of DDM based on a dual approach in steady-state analysis including the building of the interfacial problem. Then, as presented by Combescure et al. we extend this formulation to multiple time-step transient analysis. The Euler semi-implicit scheme used to integrate first-order systems and the Newmark scheme used to integrate second order systems are detailed for both single time-step and multiple time-step computations. Then, we present a discussion concerning the choice of the continuity at the interface for first-order systems in single time-step and multiple time-step transient analysis. In the second part, the numerical implementation of the proposed approach in the FEM code DynELA is presented with a focus on the treatment of multi-physic aspects. Parallelization of the code on a Shared Memory Processors (SMPs) computer is achieved using the OpenMP programming standard. The last part of this work concerns the study of an industrial benchmark concerning the power converters: the electro-thermal simulation of a semi-conductor chip in transient analysis. This example allows to compare different strategies of tearing into subdomains and to couple different time-steps on the same structure.

International audience; This article presents an original method to validate the ef- ficiency of the short length maximum separable distance forward error codes when using broadcasting mode in 802.11b/g multimedia streaming. Our method is based on real traces on which the forward error code will be mapped. Then, the quality of the streaming is evaluated by calculating the Peak Signal-to-Noise Ratio (PSNR).

International audience; In this paper we study linear conservative systems of finite dimension coupled with an infinite dimensional system of diffusive type. Computing the time-derivative of an appropriate energy functional along the solutions helps us to prove the well-posedness of the system and a stability property. But in order to prove asymptotic stability we need to apply a sufficient spectral condition. We also illustrate the sharpness of this condition by exhibiting some systems for which we do not have the asymptotic property.

Composing simulations of complex systems from already existing simulation components remains a challenging issue. Motivations for composable simulation include generation of a given federation driven by operational requirements provided "on the fly". The High Level Architecture, initially developed for designing fully distributed simulations, can be considered as an interoperability standard for composing simulations from existing components. Requirements for constructing such complex simulations are quite different from those discussed for distributed simulations. Although interoperability and reusability remain essential, both high performance and availability have also to be considered to fulfill the requirements of the end user. ONERA is currently designing a High Performance / High Availability HLA Run-time Infrastructure from its open source implementation of HLA 1.3 specifications. HP-CERTI is a software package including two main components: the first one, SHM-CERTI, provides an optimized version of CERTI based on a shared memory communication scheme; the second one, Kerrighed-CERTI, allows the deployment of CERTI through the control of the Kerrighed Single System Image operating system for clusters, currently designed by IRISA. This paper describes the design of both high performance and availability Runtime Infrastructures, focusing on the architecture of SHM-CERTI. This work is carried out in the context of the COCA (High Performance Distributed Simulation and Models Reuse) Project, sponsored by the DGA/STTC (Délégation Générale pour l'Armement/Service des Stratégies Techniques et des Technologies Communes) of the French Ministry of Defense.

National audience

La modélisation et la simulation sont souvent employées par les chercheurs de l'ONERA, par les organismes de la Défense, par les industriels, pour concevoir et valider les systèmes complexes (systèmes de défense, systèmes d'information et de commandement, systèmes embarqués, etc.). Le Projet de Recherche Fédérateur (PRF)"SICODIS" s'intéresse en particulier aux simulations distribuées à événements discrets.

International audience; This paper is organized as follows. Part A presents the context of reconfiguring transitic systems and the main idea in implementing the decision step. It comprises sections 1 to 3. Section 3 presents an example that illustrates the concepts presented in the next sections. Parts B and C express the models and principles used to simulate transitic systems, the result of which will be helpful for choosing the new configuration. Part B focuses mainly on models. It comprises sections 4 to 6. Part C focuses mainly on simulation principles. It comprises sections 7 to 10.

Recent concentration of indoor livestock farming in the Highlands of the Reunion Island generates environmental risks. Livestock effluent management is often difficult due to the lack of suitable spreading areas close to the farms. However, crop farms in the Lowlands are demanding organic matter to maintain soil fertility. This is the case of the locality of Grand-Het where pig and poultry effluents are intensively produced that should be exported towards the coastal zone where large areas of sugar-cane are available. This paper presents how we are supporting Grand-Ilet stakeholders to devise territorial management strategies of their wastes. To this end, a stepwise approach is being tested. First, the supply of animal wastes and the manage ment by farmers are characterized through farm surveys and agronomic expertise. Then, the crop demand is determined by using a GIS considering a large zone surrounding Grand-Ilet onto which several constraints are applied to eliminate unsuitable spreading areas. Comparing supply and demand enables one to draw gen-eral conclusions about the strategic choices to be considered. In the case of Grand-Ilet, distances between livestock farms and crop locations, and amounts of liquid manure to be processed, led one to consider implementing a treatment plant collectively managed by farmers. To devise sustainable management strategies for such a trade-off, we are developing a participatory approach based on several simulation models (spread-sheet, hybrid dynamical system, agent-based). These models, developed in our team, aim at tackling several decision-making issues: (i) Which treatment process should be chosen? (ii) How to supply it with raw wastes? (iii) How to best organise organic material fluxes amongst farms at the district level? Together with the description of those tools, each step of the approach is exemplified and the way of integrating the use of simulation models within the stakeholders' decision process is outlined.

Composing simulations of complex systems from already existing simulation components remains a challenging issue. Motivations for composable simulation include generation of a given federation driven by operational requirements provided "on the fly". The High Level Architecture, initially developed for designing fully distributed simulations, can be considered as an interoperability standard for composing simulations from existing components. Requirements for constructing such complex simulations are quite different from those discussed for distributed simulations. Although interoperability and reusability remain essential, both high performance and availability have also to be considered to fulfill the requirements of the end user. ONERA is currently designing a High Performance / High Availability HLA Run-time Infrastructure from its open source implementation of HLA 1.3 specifications. HP-CERTI is a software package including two main components: the first one, SHM-CERTI, provides an optimized version of CERTI based on a shared memory communication scheme; the second one, Kerrighed-CERTI, allows the deployment of CERTI through the control of the Kerrighed Single System Image operating system for clusters, currently designed by IRISA. This paper describes the design of both high performance and availability Runtime Infrastructures, focusing on the architecture of SHM-CERTI. This work is carried out in the context of the COCA (High Performance Distributed Simulation and Models Reuse) Project, sponsored by the DGA/STTC (Délégation Générale pour l'Armement/Service des Stratégies Techniques et des Technologies Communes) of the French Ministry of Defense.

This paper deals with the numerical simulation of acoustic wave propagation in axisymmetric waveguides with varying cross-section using a Webster-Lokshin model. Splitting the pipe into pieces on which the model coefficients are nearly constant, analytical solutions are derived in the Laplace domain, enabling for the realization of the propagation by concatenating scattering matrices of transfer functions (§2). These functions involve standard differential and delay operators, as well as pseudo-differential operators of diffusive type, induced by both the viscothermal losses and the curvature. These operators are explicitly decomposed thanks to an asymptotic expansion, and the diffusive ones may be defined and classified (§3). Various equivalent diffusive realizations may be proposed, that are deeply linked to choices of cuts in the complex analysis of the transfer functions. Then, finite order approximations are given for their simulation (§4).

In this paper, we discuss the design and implementation of a HLA inter-federation bridge. Our works are mainly motivated by the scalability and security problems, but we also consider the use of bridges for interoperability purposes. We describe several bridge topologies, including linear and cyclic inter-federations. We discuss problems raised by bridge federates and the use of different RTI implementations. We detail several solutions, leading to the design and implementation of a bridge prototype. Then we present our current results, and on-going works concerning performance improvements, interoperability, and security purposes.

Cette thèse s'intéresse à l'ingénierie des exigences (IE), qui caractérise le processus qui conduit à la construction d'un ensemble cohérent de spécifications portant sur un produit à construire. Nous avons identifié trois phases distinctes dans le processus d'IE : modélisation des exigences, gestion des incohérences et distribution des exigences. Dans la phase de modélisation, nous nous sommes appuyés sur une logique de préférences, CO, qui nous a permis d'exprimer les exigences de chaque agent de façon ordonnée, mais également des contraintes du domaine et des phrases normatives complexes. Nous avons alors pu définir une notion de cohérence entre ces trois notions. Pour la résolution des éventuels conflits entre exigences émises par différents agents, nous avons développé M F, une logique modale permettant de raisonner sur le contenu de bases de croyances obtenus par fusion majoritaire de plusieurs bases primitives ainsi qu'un démonstrateur automatique en PROLOG pour M F. Dans un second temps, nous avons montré que notre approche permettait de raisonner sur des ensembles d'exigences ordonnées ou non. Enfin, nous avons proposé d'inclure dans le processus d'IE une phase de distribution des exigences à un ensemble d'agents exécutants. Après avoir proposé un modèle d'agent simple, nous avons montré que ce modèle permettait de déterminer les buts d'un agent. Nous avons étendu cette approche à un groupe d'agents, puis nous avons défini un modèle de distribution comportant une entité centrale connaissant les engagements des agents. This thesis deals with requirements engineering RE characterizes the process leading to a consistent set of specifications about some product. We have identified three distinct phases in RE process : requirements modelling, inconsistency management and requirements distribution. In the modelling phase, we have used CO, a logic of preferences, which has allowed us to express each agent’s requirements in an ordonned way, but also domain constraints and complex normative sentences. We have then defined the notion of consistency between those three notions. Concerning the possible conflicts between requirements emitted by different agents, we have developped MF, a modal logic allowing to reason on belief bases obtained by majority merging. We have also developped a PROLOG automatic prover for IVIF. I/Ve have then shown that our approach allows to reason on ordonned or unordonned requirements sets. Finally, we have proposed to include in the RE process a distribution phase. The requirements are distributed among a set of executive agents. We have defined a simple model of agency from which we can derive the agent’s goals. We have then extended this approach to multiagents systems and defined a distribution model based on a central entity controlling the distribution process.

Using a particular construction of generator matrices of the q-ary image of q/sup m/-ary cyclic codes, it is proved that some of these codes are invariant under the action of particular permutation groups. The equivalence of such codes with some two-dimensional (2-D) Abelian codes and cyclic codes is deduced from this property. These permutations are also used in the area of the soft-decision decoding of some expanded Reed-Solomon (RS) codes to improve the performance of generalized minimum-distance decoding.

A new statistical approach using functions based on the circular code classifies correctly more than 93 % of bases in protein (coding) genes and non-coding genes of human sequences. Based on this statistical study, a research software called "Analysis of Coding Genes" (ACG) has been developed for identifying protein genes in the genomes and for determining their frame. Furthermore, the software ACG also allows an evaluation of the length of protein genes, their position in the genome, their relative position between themselves, and the prediction of internal frames in protein genes.

With distributed simulation, existing simulations, even from different disciplines, can be reused or made to interoperate. The efficiency of this technique is however not firmly established and it depends on the ability to satisfy a number of requirements, especially concerning simulation performance. The performance depends to a large extent on the structure and scaling of architectural components of the simulation execution platform. In this paper, we present an approach to address this problem: the design of models to capture the main characteristics of distributed simulations, run-time infrastructures and network architectures, and the development of a tool to predict performances. This tool, which is a simulator of HLA simulations is written by using HLA, and is evaluated on three test applications.

The stability of non-linear fractional differential equations is studied. A sufficient stability condition on the non-linearity is given for the input-output stability, thanks to many different reformulations of the system using diffusive representations of dissipative pseudo-differential operators. The problem of asymptotic internal stability is analyzed by a more involved functional analytic method. Finally, a fractional version of the classical Hartman-Grobman theorem for hyperbolic dynamical systems of order 1 is conjectured and reformulated, based upon known necessary and sufficient stability conditions for linear fractional differential equations.

International audience; A time-domain model for the flexural vibrations of damped plates was presented in a companion paper [Part I, J. Acoust. Soc. Am. 109, 1422-1432 (2001)]. In this paper (Part II), the damped-plate model is extended to impact excitation, using Hertz’s law of contact, and is solved numerically in order to synthesize sounds. The numerical method is based on the use of a finite-difference scheme of second order in time and fourth order in space. As a consequence of the damping terms, the stability and dispersion properties of this scheme are modified, compared to the undamped case. The numerical model is used for the time-domain simulation of vibrations and sounds produced by impact on isotropic and orthotropic plates made of various materials (aluminum, glass, carbon fiber and wood). The efficiency of the method is validated by comparisons with analytical and experimental data. The sounds produced show a high degree of similarity with real sounds and allow a clear recognition of each constitutive material of the plate without ambiguity.

A large class of visco-elastic and elasto-plastic systems, frequently encountered in physics, are based on causal pseudo-differential operators, which are hereditary: the whole past of the state is involved in the dynamic expression of the system evolution. This generally induces major technical difficulties. We consider a specific class of pseudo-differential damping operators, associated to the so-called diffusive representation which enables to built augmented state-space realizations without heredity. Dissipativity property is expressed in a straightforward and precise way. Thanks to state-space realizations, standard analysis and approximation methods as well as control-theory concepts may therefore be used.

The High Level Architecture (HLA) establishes a common modeling and simulation framework facilitating interoperability and reuse of simulation components. Since 1996, ONERA (French Aeronautics and Space Research Centre) carries out several studies on HLA in order to gain a better understanding of the underlying mechanisms of HLA implementations. The first critical step of this initiative was to develop our own RTI from the HLA specifications. In order to evaluate the cost of making a transition from legacy simulations to HLA, we first developed an HLA federation simulating an air-ground combat involving a set of aircraft's engaged against a surface to air defense system. Current studies on HLA distributed simulation include security, WAN simulations and multiresolution. Conventional simulations represent entities at just one single level of resolution. Multiresolution representation of entities consists in maintaining multiple and concurrent representations of entities. In this paper we address the problem of how HLA services may allow to achieve multiresolution modeling and simulation. Our goal is not to provide a general framework as a basis for designing simulations of entities at different levels of resolution concurrently. We focus on experience feedback we have obtained by migrating a single level resolution HLA federation to a multi-level resolution federation. The selected application is an air-ground combat simulation involving aggregated patrols of aircraft's engaged against a surface to air defense system. In this paper, we briefly describe the air-ground combat simulation application. We then detail the multiresolution representation of entities (patrols and aircraft's), and discuss the chosen mechanisms allowing triggering aggregation from an entity-level representation, and conversely, triggering disaggregation from an aggregate representation. We focus on the HLA services we have selected to maintain several levels of representation concurrently and on methodological issues in designing multiresolution HLA simulations. We have tackled some difficulties and we propose a new HLA service that should make easier the user's task. This multiresolution management service can be added to our RTI or written by using existing HLA services. Finally, future trends are discussed.