The amount of information on the World Wide Web (WWW) is rapidly growing in pace and topic diversity. This has made it increasingly difficult, and often frustrating, for information seekers to retrieve the content they are looking for as information retrieval systems (e.g., search engines) are unable to decipher the relevance of the retrieved information as it pertains to the information they are searching for. This issue can be decomposed into two aspects: 1) variability of information relevance as it pertains to an information seeker. In other words, different information seekers may enter the same search text, or keywords, but expect completely different results. It is therefore, imperative that information retrieval systems possess an ability to incorporate a model of the information seeker in order to estimate the relevance and context of use of information before presenting results. Of course, in this context, by a model we mean the capture of trends in the information seeker's search behaviour. This is what many researchers refer to as the personalized search. 2) Information diversity. Information available on the World Wide Web today spans multitudes of inherently overlapping topics, and it is difficult for any information retrieval system to decide effectively on the relevance of the information retrieved in response to an information seeker's query. For example, the information seeker who wishes to use WWW to learn about a cure for a certain illness would receive a more relevant answer if the search engine was optimized into such domains of topics. This is what is being referred to in the WWW nomenclature as a 'specialized search'. This thesis maintains that the information seeker's search is not intended to be completely random and therefore tends to portray itself as consistent patterns of behaviour. Nonetheless, this behaviour, despite being consistent, can be quite complex to capture. To accomplish this goal the thesis proposes a Multi-Agent Personalized Information Retrieval with Specialization Ontology (MAPIRSO). MAPIRSO offers a complete learning framework that is able to model the end user's search behaviour and interests and to organize information into categorized domains so as to ensure maximum relevance of its responses as they pertain to the end user queries. Specialization and personalization are accomplished using a group of collaborative agents. Each agent employs a Reinforcement Learning (RL) strategy to capture end user's behaviour and interests. Reinforcement learning allows the agents to evolve their knowledge of the end user behaviour and interests as they function to serve him or her. Furthermore, REL allows each agent to adapt to changes in an end user's behaviour and interests. Specialization is the process by which new information domains are created based on existing information topics, allowing new kinds of content to be built exclusively for information seekers. One of the key characteristics of specialization domains is the seeker centric - which allows intelligent agents to create new information based on the information seekers' feedback and their behaviours. Specialized domains are created by intelligent agents that collect information from a specific domain topic. The task of these specialized agents is to map the user's query to a repository of specific domains in order to present users with relevant information. As a result, mapping users' queries to only relevant information is one of the fundamental challenges in Artificial Intelligent (AI) and machine learning research. Our approach employs intelligent cooperative agents that specialize in building personalized ontology information domains that pertain to each information seeker's specific needs. Specializing and categorizing information into unique domains is one of the challenge areas that have been addressed and various proposed solutions were evaluated and adopted to address growing information. However, categorizing information into unique domains does not satisfy each individualized information seeker. Information seekers might search for similar topics, but each would have different interests. For example, medical information of a specific medical domain has different importance to both the doctor and patients. The thesis presents a novel solution that will resolve the growing and diverse information by building seeker centric specialized information domains that are personalized through the information seekers' feedback and behaviours. To address this challenge, the research examines the fundamental components that constitute the specialized agent: an intelligent machine learning system, user input queries, an intelligent agent, and information resources constructed through specialized domains. Experimental work is reported to demonstrate the efficiency of the proposed solution in addressing the overlapping information growth. The experimental work utilizes extensive user-centric specialized domain topics. This work employs personalized and collaborative multi learning agents and ontology techniques thereby enriching the queries and domains of the user. Therefore, experiments and results have shown that building specialized ontology domains, pertinent to the information seekers' needs, are more precise and efficient compared to other information retrieval applications and existing search engines.

La disponibilité d'images satellites à très haute résolution spatiale au dessus de zones urbaines est récente. Elle constitue potentiellement un très grand apport pour la cartographie des villes à des échelles de l'ordre du 1:10 000. La très haute résolution spatiale permet une représentation réelle des rues sue une carte, mais engendre une augmentation significative du bruit. Dans cette thèse, nous proposons une méthode d'extraction des réseaux de rues en milieu urbain à partir des images à très haute résolution spatiale. Son objectif est de répondre à une forte demande dans la création automatisée de cartes. La méthode proposée n'utilise que l'image numérique comme source d'information. Elle est semi-automatique au niveau de la détection et exploite la coopération entre la représentation linéique de la rue et sa représentation surfacique. Le graphe topologique du réseau est d'abord extrait et est utilisé pour initialiser l'étape de reconstruction surfacique. Le résultat d'extraction peut alors servir à recaler le graphe précisément sur l'axe des rues. La méthode utilise des contraintes géométriques fortes afin de ne pas dépendre d'un modèle de profil radiométrique de la rue, trop variable en milieu urbain. Dans cette optique, un modèle de contours actif associé à la transformée en ondelettes, le DoubleSnake, a été développé. Son évolution dans un cadre multi-échelle permet d'extraire les sections de rues à bords parallèles dans un environnement bruité. Les positions finales des DoubleSnakes permettent ensuite l'extraction des intersections. La méthode a été appliquée à des images de différents capteurs et avec différents types d'urbanisation. Un protocole innovant d'évaluation quantitative des résultats par comparaison à l'interprétation humaine a permis de montrer le caractère générique de la méthode, ainsi que sa bonne robustesse vis-à-vis du bruit. Cette méthode constitue un pas vers une cartographie automatisée du réseau de rues urbain.

A bord des véhicules électriques (VE) et Hybrides (VEH), les fonctions de tractions sont assurées par des convertisseurs électroniques de puissances. Ces derniers sont constitués de module de puissance (IGBTs ou MOSFETs). Au cours de leur fonctionnement, ces modules sont parfois soumis à de fortes contraintes électriques et thermiques qui amènent à une défaillance ou même à une destruction. Le premier objectif sera de réaliser un banc expérimentale permettant d'étudier le vieillissement des modules IGBTs en régîmes extrêmes de fonctionnement (mode de court-circuit). Ainsi, nous évaluerons les différents indicateurs de vieillissements permettant de prédire la défaillance du composant. Il sera question aussi de suivre le vieillissement ou une dégradation initié sur les composants IGBTs par thermographie infrarouge. Le second objectif sera de modéliser et simuler par éléments finis différentes structures d'IGBTs, afin de valider les modèles en fonctionnement statique et dynamique. L'avantage de l'approche multicellulaire par rapport à l'approche unicellulaire sera mis en avant.

The amount of information on the World Wide Web (WWW) is rapidly growing in pace and topic diversity. This has made it increasingly difficult, and often frustrating, for information seekers to retrieve the content they are looking for as information retrieval systems (e.g., search engines) are unable to decipher the relevance of the retrieved information as it pertains to the information they are searching for. This issue can be decomposed into two aspects: 1) variability of information relevance as it pertains to an information seeker. In other words, different information seekers may enter the same search text, or keywords, but expect completely different results. It is therefore, imperative that information retrieval systems possess an ability to incorporate a model of the information seeker in order to estimate the relevance and context of use of information before presenting results. Of course, in this context, by a model we mean the capture of trends in the information seeker's search behaviour. This is what many researchers refer to as the personalized search. 2) Information diversity. Information available on the World Wide Web today spans multitudes of inherently overlapping topics, and it is difficult for any information retrieval system to decide effectively on the relevance of the information retrieved in response to an information seeker's query. For example, the information seeker who wishes to use WWW to learn about a cure for a certain illness would receive a more relevant answer if the search engine was optimized into such domains of topics. This is what is being referred to in the WWW nomenclature as a 'specialized search'. This thesis maintains that the information seeker's search is not intended to be completely random and therefore tends to portray itself as consistent patterns of behaviour. Nonetheless, this behaviour, despite being consistent, can be quite complex to capture. To accomplish this goal the thesis proposes a Multi-Agent Personalized Information Retrieval with Specialization Ontology (MAPIRSO). MAPIRSO offers a complete learning framework that is able to model the end user's search behaviour and interests and to organize information into categorized domains so as to ensure maximum relevance of its responses as they pertain to the end user queries. Specialization and personalization are accomplished using a group of collaborative agents. Each agent employs a Reinforcement Learning (RL) strategy to capture end user's behaviour and interests. Reinforcement learning allows the agents to evolve their knowledge of the end user behaviour and interests as they function to serve him or her. Furthermore, REL allows each agent to adapt to changes in an end user's behaviour and interests. Specialization is the process by which new information domains are created based on existing information topics, allowing new kinds of content to be built exclusively for information seekers. One of the key characteristics of specialization domains is the seeker centric - which allows intelligent agents to create new information based on the information seekers' feedback and their behaviours. Specialized domains are created by intelligent agents that collect information from a specific domain topic. The task of these specialized agents is to map the user's query to a repository of specific domains in order to present users with relevant information. As a result, mapping users' queries to only relevant information is one of the fundamental challenges in Artificial Intelligent (AI) and machine learning research. Our approach employs intelligent cooperative agents that specialize in building personalized ontology information domains that pertain to each information seeker's specific needs. Specializing and categorizing information into unique domains is one of the challenge areas that have been addressed and various proposed solutions were evaluated and adopted to address growing information. However, categorizing information into unique domains does not satisfy each individualized information seeker. Information seekers might search for similar topics, but each would have different interests. For example, medical information of a specific medical domain has different importance to both the doctor and patients. The thesis presents a novel solution that will resolve the growing and diverse information by building seeker centric specialized information domains that are personalized through the information seekers' feedback and behaviours. To address this challenge, the research examines the fundamental components that constitute the specialized agent: an intelligent machine learning system, user input queries, an intelligent agent, and information resources constructed through specialized domains. Experimental work is reported to demonstrate the efficiency of the proposed solution in addressing the overlapping information growth. The experimental work utilizes extensive user-centric specialized domain topics. This work employs personalized and collaborative multi learning agents and ontology techniques thereby enriching the queries and domains of the user. Therefore, experiments and results have shown that building specialized ontology domains, pertinent to the information seekers' needs, are more precise and efficient compared to other information retrieval applications and existing search engines.

Cette thèse étudie l’organisation longitudinale du lit du Rhône français et son évolution depuis plus d’un siècle afin d’améliorer la gestion des sédiments du fleuve. Elle décrit d’abord la complexité géographique, géologique, et hydrologique des 512 km du système fluvial. Les extractions, l’endiguement, la chenalisation et les aménagements hydroélectriques ont simplifié le tracé latéral du chenal, incisé et pavé le fond du lit et appauvri les habitats aquatiques et riverains. La thèse caractérise ensuite plus précisément les dynamiques de fond. Elle repose sur des données bathymétriques de 1897-2010 et sur des données granulométriques issues de plus de 300 prélèvements réalisés dans le cadre de ces travaux sur l’ensemble du linéaire. Un protocole d’échantillonnage homogène a été spécialement élaboré. Les variations altimétriques du lit (zones de stockage-érosion et faciès morphologiques) et sa composition granulométrique (zones pavées ou comblées par des fines) ont été caractérisées afin d’évaluer la mobilité potentielle de la charge de fond, les apports des affluents et l’impact des ouvrages sur les conditions de transport. Il ressort 1) que l’incision du lit est principalement due à la chenalisation, 2) que les aménagements hydroélectriques ont compartimenté le continuum en provoquant le stockage de sédiments fins dans certaines retenues et en favorisant le pavage dans les vieux Rhône à l’aval de Lyon et 3) qu’une dynamique résiduelle subsiste lors de crues majeures avec une remobilisation modérée de la charge de fond. Les données des secteurs les plus pavés ont été intégrées dans un modèle hydraulique 1D permettant de quantifier la capacité de transport et le débit solide et confirmant les résultats précédents. Ces travaux permettent de formuler des recommandations de gestion durable et de restauration écologique du fleuve telles que l’ouverture des vannes de fond en crue, la modification des consignes aux barrages ou la réinjection de graviers. / This thesis studies the French Rhone riverbed longitudinal organisation and its evolution over a century, with a view to improve the management strategy for the sediments. It first describes the geographic, geologic and hydrologic peculiarities of the 512-km long fluvial system. Gravel mining, dikes construction, channelization and dams construction have simplified the lateral pattern of the channel, incised and paved the riverbed and impoverished the aquatic and riparian habitats. The thesis then more precisely characterizes the riverbed dynamics. It relies on bathymetric data collected since 1897 and on grainsize data collected specifically for this study from over 300 samples on the whole 512-km river length. An homogeneous sampling protocol was elaborated on purpose. The vertical bed evolution and grainsize distribution were analysed to assess the riverbed potential mobility, tributaries’ sediment recharge and infrastructures’ impact on sediment transport. It thus appears that 1) riverbed incision is mainly due to channelization, 2) hydroelectric dams partitioned the continuum with the accumulation of fine sediments in some backwaters and the pavement of bypassed reaches downstream of Lyon and 3) a residual dynamic persists during major floods, with moderate gravels transport. The data of paved reaches were used in a 1D hydraulic model to quantify transport capacity and bedload discharge, confirming the aforementioned results. This study helps formulate such recommendations for sustainable river management and ecological restoration as opening slush gates during floods, changing dams management instructions or reinjecting gravel.

A bord des véhicules électriques (VE) et Hybrides (VEH), les fonctions de tractions sont assurées par des convertisseurs électroniques de puissances. Ces derniers sont constitués de module de puissance (IGBTs ou MOSFETs). Au cours de leur fonctionnement, ces modules sont parfois soumis à de fortes contraintes électriques et thermiques qui amènent à une défaillance ou même à une destruction. Le premier objectif sera de réaliser un banc expérimentale permettant d’étudier le vieillissement des modules IGBTs en régîmes extrêmes de fonctionnement (mode de court-circuit). Ainsi, nous évaluerons les différents indicateurs de vieillissements permettant de prédire la défaillance du composant. Il sera question aussi de suivre le vieillissement ou une dégradation initié sur les composants IGBTs par thermographie infrarouge. Le second objectif sera de modéliser et simuler par éléments finis différentes structures d’IGBTs, afin de valider les modèles en fonctionnement statique et dynamique. L’avantage de l’approche multicellulaire par rapport à l’approche unicellulaire sera mis en avant. / On board electric vehicles (EVs) and hybrid (HEV), the functions of traction is provided by power electronic converters. These consist of power modules (IGBT or MOSFET). During their operation, these modules are sometimes subjected to high electrical and thermal stresses that lead to failure or even destruction.The first objective will be to achieve experimental bench to study ageing IGBT modules under extreme operating conditions ( short circuit mode). Thus, we evaluate the various indicators of ageing to predict component failure. Topics will also follow the ageing or degradation initiated on IGBT components by infrared thermography. The second objective is to model and simulate by finite element different IGBT structures to validate the models in static and dynamic operation. The advantage of multicellular approach to the single cell approach will be highlighted.

Pharmaceuticals are biological active compounds with different functional groups, physico-chemical and biological properties. These chemical compounds are called as “new contaminants” which cumulate in various environmental components. These contaminants input to environment from industrial processes, hospitals and health care institutions or household sources and these pollutants can to negatively interact with environmental components. Pharmaceuticals are separated according to structure and their effects to live organisms. On the basis of drug consumption in Czech Republic the non-steroidal anti-inflammatory drugs are the most using pharmaceuticals, because these drugs can be obtain without prescriptions. The choosing analgesics for study were monitored in waste water from waste waters treatment plant (WWTP Brno – Modřice) and in surface water from two sampling sites of river Kretinka. Pharmaceuticals were monitored in surface waters in Czech Republic, in Scotland (river Thurso) and in Taiwan (river Erren, Agongdian, Yanshuei and canal of Tainan) Solid phase extraction (SPE) is used for extraction of study pharmaceutical from water system. The high performance liquid chromatography (HPLC) with diode array detection or mass spectrometry was used for determination of drugs in surface waters from Czech Republic. The water samples from Scotland and Taiwan were analysed and the high performance liquid chromatography with mass spectrometry (HPLC/MS) was used for determination of pharmaceuticals. The method of HPLC/MS enables the determination of study compounds in the very low range of concentrations (ng ? l-1). All monitored pharmaceuticals were identified and quantified in water samples from river Kretinka in Czech Republic, river Thurso (Scotland) and rivers Erren, Agongdian, Yanshuei and canal of Tainan in Taiwan and the obtained data were compared.

From the Proceedings of the 1974 Meetings of the Arizona Section - American Water Resources Assn. and the Hydrology Section - Arizona Academy of Science - April 19-20, 1974, Flagstaff, Arizona / The process of salt replacement desalination proposed is believed to be able to produce vast quantities of fresh water be desalination. This method, which is a novel approach to minimizing the costs of saline water conversion, consists of the substitution of solutes in a solution to be desalted by a replacer chemical, and the low energy removal of that replacer chemical. The ultrafiltration of larger molecular sized replacer chemicals with high flux membranes increases the produce yield rate and reduces the corresponding energy requirement, with respect to reverse osmosis. In addition, the initial captial investment is less since no pressure constraining devices are required. The alteration of the osmotic pressure of the replacer solution within the process can also take advantage of energy savings through the utilization of an easily reversible reaction which synthesizes and breaks down a constituent that has a significant osmotic pressure difference between phases. Finally, the unusual process of fixed gel syneresis shows potential as a low energy salt replacement type process, but still requires extensive investigation.

Depuis au moins 2000 ans, le massif des Vosges a été le siège d’activités minières qui ont conduit (i) à l’émission de particules contaminées en éléments traces métalliques (ETMs), (ii) aux dépôts de déchets miniers et sidérurgiques et (iii) à des modifications du paysage. Les ressources présentes dans ce système de moyenne montagne, et utilisées par les populations, rendent nécessaire la mesure des impacts des activités anthropiques passées sur l’écosystème forestier.Ce travail de thèse propose une approche multi-proxy couplant géochimie, palynologie et écotoxicologie. Différentes phases de dépôts en ETMs ont été retracées sur les 2000 dernières années et regroupent celles connues à l’échelle de l’hémisphère Nord et des phases plus locales. L’impact paysager des activités minières semble quant à lui réduit par rapport à ce que laissaient penser les archives historiques qui présentaient des déserts forestiers dus aux activités minières. La biodisponibilité et le risque associés au transfert des ETMs ont été évalués dans un ancien district minier de production de plomb (Pb) et d’argent grâce à une campagne de bioindication active avec un indicateur de la qualité des sols, Cantareus aspersus. Sur huit stations, quatre présentent des excès de transfert induisant un risque toxicologique élevé et comprennent deux sites archéologiques miniers et un potager communal. Le fractionnement et la spéciation du Pb permettent de montrer une pratique à risque avec l’utilisation de scories comme amendements sur le potager. La contamination du milieu en ETMs par les activités minières est ici toujours transférable aux organismes du sol et donc potentiellement dans le réseau trophique / Since at least the last two millennia, mining and smelting activities have occurred in the Vosges Mountains and conducted to (i) the emission of trace metal (TM) contaminated particles in the atmosphere, (ii) the presence of mining and smelting wastes and (iii) landscape changes. Due to the use of resources from this mountainous area by local populations, it is necessary to measure impacts of these past activities on the forested ecosystem.This PhD thesis proposes a multi-proxy approach coupling geochemistry, palynology and ecotoxicology. Several phases of TM deposits have been recorded and correspond to major phases of TM contamination known in the northern hemisphere and also local periods of mining/smelting activities. The impact of mining activities on the landscape appears to be reduced contrary to historical data that described shortage of wood due to mining activities. The bioavailability and the risk associated to TMs from mining activities have been assessed in a former lead (Pb)-silver mining district by an active bioindication campaign using Cantareus aspersus, considered as indicator of soil quality. Among the eight stations, four of them present an excess of TM transfer conducting to a high ecological risk, particularly in two archaeological mining sites but also in a kitchen garden. Lead fractionation and speciation highlight the use of slags as amendments in the kitchen garden. Trace metal contamination of the soil due to mining activities is still bioavailable for soil organisms and so potentially transferred in food chain.