Que savons-nous de la didactisation des pratiques de savoir scientifiques dans le contexte d’un événement public comme la Fête de la science ? Notre recherche étudie ces questions ; nous avons retenu comme terrain d’enquête trois stands d’un village des sciences, à propos des os et articulations, des eaux souterraines et du sable. Les ateliers sont menés par des scientifiques ; le public sur lequel nous enquêtons est scolaire ou non scolaire. Nous menons nos enquêtes auprès des chercheurs dans leur laboratoire pendant la préparation de l’événement, mais aussi dans leur activité ordinaire. Nous suivons une classe d’école élémentaire avant, pendant et après l’événement. Nos cadres théoriques sont la théorie de l’action conjointe en didactique, et l’approche documentaire. Notre méthode qualitative est basée sur des études de cas et s’appuie sur des films d’étude. Nous développons une méthodologie adaptée à cette recherche, sur la base de traitements d’images et de curation des données. Nous donnons à voir nos descriptions, nos analyses et nos résultats avec des systèmes hybrides texte-image-son. Nous proposons une typologie des transactions didactiques en situation de travail conjoint de plusieurs professeurs avec le même groupe d’élèves. Dans ce cadre spécifique, nous contribuons à la modélisation de certaines notions clés : le triplet des genèses (topogenèse, mésogenèse et chronogenèse), les transactions didactiques, les sémioses. Notre recherche met en évidence que l’activité de didactisation des pratiques scientifiques est contiguë et parallèle à leur production. Elle est nécessaire au sein même des laboratoires et s’ajuste aux publics. L’activité de didactisation sur les stands prend en compte toute pratique de savoir. Elle vise à replacer dans leur contexte les pratiques incompatibles avec les pratiques scientifiques. Elle prend en compte les pratiques de savoir compatibles avec les pratiques scientifiques, et amène le public à voir les thématiques abordées comme les scientifiques les voient. Les artefacts utilisés par les « chercheurs-médiateurs » sont ainsi conçus pour donner à voir des phénomènes ciblés. Ils sont combinés entre eux et associés à d’autres ressources matérielles et symboliques, sémiotiques et proxémiques. Le travail empirique a été conduit dans un contexte spécifique et sur un public particulier. Nos résultats doivent être mis à l’épreuve dans d’autres situations. Les ressources méthodologiques et technologiques demandent à être développées, testées et intégrées dans le système de ressources d’autres chercheurs. / What do we know about scientific knowledge practices in the specific context of the French event « Festival of science »? The research studies this question. We have chosen to investigate three stands in a “Life sciences village” about bones and articulations, underground water and sand. Activities are being carried out by scientists; the public can come from schools, but can also be a general public. We investigate in the scientists preparation of the event in their laboratory, and we look at their ordinary activity too. We observe students of an elementary school before, during and after the event. We refer to the theoretical frameworks of the Joint Action Theory in Didactics, and the Documentational Approach. Our qualitative method is based on case studies. The central data we analyze are videos shot during the investigations. We develop a suitable methodology for this research, based on image processing and digital data curation. We provide descriptions, analysis and results with text-picture-sound hybrid systems. We map out a typology of didactic transactions in the case of joint work of teachers with the same group of students. In this specific framework, we contribute to model concepts as: the genesis triplets (topo genesis, meso genesis and chrono genesis), the didactic transactions, the semiosis process. The research highlights that the activity of didactisation of scientific practices is contiguous with and parallel to its production. It is necessary within research labs themselves and has to adjust to the public. In the stands all knowledge practices are studied: some knowledge practices, inconsistent with scientific knowledge, are replaced in their original context. The activity of didactisation associates knowledge practices (consistent with scientific knowledge) with this scientific knowledge. Its aim is to bring the views of the public closer from those of the scientists. Researchers communicate their scientific activity to a public of non-specialists with artifacts which evidence phenomena. These artefacts are linked and associated with other material and symbolic, semiotic and proxemic, resources. The empiric work had been done in a specific context and with a particular public. These results now need to be implemented in other situations. Methodological and technological resources must be further developed and tested in other research.

O presente trabalho apresenta dois novas modelos conexionistas, baseados na teoria da adaptação ressonante (ART): Simplified Fuzzy ARTMAP e Semantic ART (SMART). Descreve-se a modelagem, adaptação, implementação e validação destes, enquanto incorporados ao sistema hibrido HYCONES, para resolução de problemas de diagnostico medico em cardiopatias congênitas e nefrologia. HYCONES é uma ferramenta para a construção de sistemas especialistas híbridos que integra redes neurais com frames, assimilando as qualidades inerentes aos dois paradigmas. 0 mecanismo de frames fornece tipos construtores flexíveis para a modelagem do conhecimento do domínio, enquanto as redes neurais, representadas na versão original de HYCONES pelo modelo neural combinatório (MNC), possibilitam tanto a automação da aquisição de conhecimento, a partir de uma base de casos, quanta a implementação de aprendizado indutivo e dedutivo. A teoria da adaptação ressonante 6 caracterizada, principalmente, pela manutenção do equilíbrio entre as propriedades de plasticidade e estabilidade durante o processo de aprendizagem. ART inclui vários modelos conexionistas, tais como: Fuzzy ARTMAP, Fuzzy ART, ART 1, ART 2 e ART 3. Dentre estes, a rede neural Fuzzy ARTMAP destaca-se por possibilitar o tratamento de padr6es analógicos a partir de dois módulos ART básicos. O modelo Simplified Fuzzy ARTMAP, como o pr6prio nome o diz, a uma simplificação da rede neural Fuzzy ARTMAP. Ao contrario desta, o novo modelo possibilita o tratamento de padrões analógicos, a partir de apenas um modulo ART, responsável pelo tratamento dos padrões de entrada, adicionado de uma camada, responsável pelos padrões alvo. Mesmo com apenas um modulo ART, o modelo Simplified Fuzzy ARTMAP 6 capaz de reter o mesmo nível de desempenho obtido com a rede neural Fuzzy ARTMAP pois, continua a garantir, conjuntamente, a maximização da generalização e a minimização do erro preditivo, através da execução da estratégia match-tracking. Para a construção da base de casos de cardiopatias congênitas, 66 prontuários médicos, das três cardiopatias congênitas mais freqüentes, foram extraídos do banco de dados de pacientes submetidos a cirurgia cardíaca no Instituto de Cardiologia RS (ICFUC-RS). Tais prontuários abrangem o período de janeiro de 1986 a dezembro de 1990 e reportam 22 casos de Comunicação Interatrial (CIA), 29 de Comunicação Interventricular (CIV) e 15 de Defeito Septal Atrioventricular (DSAV). Para a análise de desempenho do sistema, 33 casos adicionais, do referido período, foram extraídos aleatoriamente do banco de dados do ICFUC-RS. Destes 33 casos, 13 apresentam CIA, 10 CIV e 10 DSAV. Para a construção da base de casos de síndromes renais, 381 prontuários do banco de dados de síndromes renais da Escola Paulista de Medicina foram analisados e 58 evidencias, correspondentes a dados de hist6ria clinica e exame físico dos pacientes, foram extraídas semi-automaticamente. Do total de casos selecionados, 136 apresentam Uremia, 85 Nefrite, 100 Hipertensão e 60 Litiase. Dos 381 casos analisados, 254 foram escolhidos aleatoriamente para a composicao do conjunto de treinamento, enquanto que os demais foram utilizados para a elaboração do conjunto de testes. Para que HYCONES II fosse validado, foram construídas 46 versões da base de conhecimento hibrida (BCH) para o domínio de cardiopatias congênitas e 46 versões da BCH para o de nefrologia. Em ambos os domínios médicos as respectivas bases de conhecimento foram construídas, automaticamente, a partir das respectivas bases de casos de treinamento. Das 46 versões geradas para cada grupo, uma representa o modelo MNC e 45 os modelos ART. As versões ART dividem-se em grupos de 3: 15 versões foram formadas a partir do modelo Simplified Fuzzy ARTMAP; 15 a partir deste mesmo modelo, sem que os padrões de entrada fossem normalizados; e, finalmente, 15 para o modelo Semantic ART. Na base de testes CHD, o desempenho da versa° HYCONES II - Simplified Fuzzy ARTMAP foi semelhante ao da versa° MNC. A primeira acertou 29 dos 33 diagnósticos (87,9%), enquanto a segunda apontou corretamente 31 dos 33 diagnósticos apresentados (93,9%). Na base de testes de síndromes renais, o desempenho de HYCONES II Fuzzy ARTMAP foi superior ao da versão MNC (p < 0,05). Ambas -Simplified acertaram, respectivamente, 108 (85%) e 95 (74,8%) diagnósticos, em 127 casos submetidos. Ainda que o desempenho da versão HYCONES II - Simplified Fuzzy ARTMAP se revelasse promissor, ao se examinar o conteúdo das redes geradas por este modelo, pode-se observar que estas divergiam completamente daquelas obtidas pelo MNC. As redes que levaram a conclusão diagnostica, na versão HYCONES - MNC, possuíam conteúdo praticamente igual aos grafos de conhecimento, elicitados de especialistas em cardiopatias congênitas. JA, as redes ativadas na versa° HYCONES II - Simplified Fuzzy ARTMAP, além de representarem numero bem major de evidencias que as redes MNC, a grande maioria destas ultimas representam a negação do padrão de entrada. Este fato deve-se a um processo de normalização, inerente ao modelo Simplified Fuzzy ARTMAP, no qual cada padrão de entrada e duplicado. Nesta duplicação, são representadas as evidências presentes em cada caso e, ao mesmo tempo, complementarmente, as evidencias ausentes, em relação ao total geral das mesmas na base de casos. Esta codificação inviabiliza o mecanismo de explanação do sistema HYCONES, pois, na área módica, os diagnósticos costumam ser feitos a partir de um conjunto de evidencias presentes e, não, pela ausência delas. Tentou-se, então, melhorar o conteúdo semântico das redes Simplified Fuzzy ARTMAP. Para tal, o processo de normalização ou codificação complementar da implementação do modelo foi retirado, validando-o novamente, contra o mesma base de testes. Na base de testes CHD, o desempenho de HYCONES II - Simplified Fuzzy ARTMAP, sem a codificação complementar, foi inferior ao da versão MNC (p < 0,05). A primeira acertou 25 dos 33 diagnósticos (75,8%), enquanto a segunda apontou corretamente 31 dos mesmos (93,9%). Na base de testes renais, o desempenho da versa° HYCONES II - Simplified Fuzzy ARTMAP, sem a codificação complementar, foi semelhante ao da versa° MNC. Dos 127 casos apresentados, a primeira acertou 98 diagn6sticos (77,2%), contra 95 da segunda (74,8%). Constatou-se, ainda, que as categorias de reconhecimento formadas pelo modelo Simplified Fuzzy ARTMAP continuavam a apresentar diferenças marcantes quanto ao seu conteúdo, quando comparadas as redes MNC ou aos grafos de conhecimento elicitados de especialistas. O modelo Semantic ART foi, então, proposto, na tentativa de se melhorar o conteúdo semantic° das redes ART. Modificou-se, então, o algoritmo de aprendizado do modelo Simplified Fuzzy ARTMAP, introduzindo-se o mecanismo de aprendizado indutivo do modelo MNC, i.e., o algoritmo de punições e recompensas, associado ao de poda e normalização. Nova validação com a mesma base de testes foi realizada. Para a base de testes de CHD, o desempenho de HYCONES II - SMART foi semelhante ao da versão Simplified Fuzzy ARTMAP e da versão MNC. A primeira e a segunda acertaram 29 dos 33 diagnósticos (87,9%), enquanto a versão MNC apontou corretamente 31 dos 33 diagnósticos apresentados (93,9%). Na base de testes de síndromes renais, o desempenho de HYCONES II - SMART foi superior ao da versão MNC (p < 0,05) e igual ao da versão Simplified Fuzzy ARTMAP. A primeira e a Ultima acertaram 108 dos 127 diagnósticos (85%), enquanto a segunda apontou corretamente 95 dos mesmos (74,8%). Desta feita, observou-se que as redes neurais geradas por HYCONES II - SMART eram semelhantes em conteúdo as redes MNC e aos grafos de conhecimento elicitados de múltiplos especialistas. As principais contribuições desta dissertação são: o projeto, implementação e validação dos modelos Simplified Fuzzy ARTMAP e SMART. Destaca-se, porem, o modelo SMART, que apresentou major valor semântico nas categorias de reconhecimento do que o observado nos modelos ART convencionais, graças a incorporação dos conceitos de especificidade e relevância. Esta dissertação, entretanto, representa não só a modelagem e validação de dois novos modelos neurais, mas sim, o enriquecimento do sistema HYCONES, a partir da continuação de dissertação de mestrado previamente defendida. A partir do presente trabalho, portanto, é dada a possibilidade de escolha, ao engenheiro de conhecimento, de um entre três modelos neurais: o MNC, o Semantic ART e o Simplified Fuzzy ARTMAP que, sem exceção, apresentam Born desempenho. Os dois primeiros destacam-se, contudo, por suportarem semanticamente o contexto. / This dissertation presents two new connectionist models based on the adaptive resonance theory (ART): Simplified Fuzzy ARTMAP and Semantic ART (SMART). The modeling, adaptation, implementation and validation of these models are described, in their association to HYCONES, a hybrid connectionist expert system to solve classification problems. HYCONES integrates the knowledge representation mechanism of frames with neural networks, incorporating the inherent qualities of the two paradigms. While the frames mechanism provides flexible constructs for modeling the domain knowledge, neural networks, implemented in HYCONES' first version by the combinatorial neuron model (CNM), provide the means for automatic knowledge acquisition from a case database, enabling, as well, the implementation of deductive and inductive learning. The Adaptive Resonance Theory (ART) deals with a system involving selfstabilizing input patterns into recognition categories, while maintaining a balance between the properties of plasticity and stability. ART includes a series of different connectionist models: Fuzzy ARTMAP, Fuzzy ART, ART 1, ART 2, and ART 3. Among them, the Fuzzy ARTMAP one stands out for being capable of learning analogical patterns, using two basic ART modules. The Simplified Fuzzy ARTMAP model is a simplification of the Fuzzy ARTMAP neural network. Constrating the first model, the new one is capable of learning analogical patterns using only one ART module. This module is responsible for the categorization of the input patterns. However, it has one more layer, which is responsible for receiving and propagating the target patterns through the network. The presence of a single ART module does not hamper the Simplified Fuzzy ARTMAP model. The same performance levels are attained when the latter one runs without the second ART module. This is certified by the match-tracking strategy, that conjointly maximizes generalization and minimizes predictive error. Two medical domains were chosen to validate HYCONES performance: congenital heart diseases (CHD) and renal syndromes. To build up the CHD case base, 66 medical records were extracted from the cardiac surgery database of the Institute of Cardiology RS (ICFUC-RS). These records cover the period from January 1986 to December 1990 and describe 22 cases of Atrial Septal Defect (ASD), 29 of Ventriculal Septal Defect (VSD), and 15 of Atrial- Ventricular Septa! Defect (AVSD), the three most frequent congenital heart diseases. For validation purposes, 33 additional cases, from the same database and period mentioned above, were also extracted. From these cases, 13 report ASD, 10 VSD and 10 AVSD. To build the renal syndromes case base, 381 medical records from the database of the Escola Paulista de Medicina were analyzed and 58 evidences, covering the patients' clinical history and physical examination data, were semiautomatically extracted. From the total number of selected cases, 136 exhibit Uremia, 85 Nephritis, 100 Hypertension, and 60 Calculosis. From the 381 cases analyzed, 245 were randomically chosen to build the training set, while the remaining ones were used to build the testing set. To validate HYCONES II, 46 versions of the hybrid knowledge base (HKB) with congenital heart diseases were built; for the renal domain, another set of 46 HKB versions were constructed. For both medical domains, the HKBs were automatically generated from the training databases. From these 46 versions, one operates with the CNM model and the other 45 deals with two ART models. These ART versions are divided in three groups: 15 versions were built using the Simplified Fuzzy ARTMAP model; 15 used the Simplified Fuzzy ARTMAP model without the normalization of the input patterns, and 15 used the Semantic ART model. HYCONES II - Simplified Fuzzy ARTMAP and HYCONES - CNM performed similarly for the CH D domain. The first one pointed out correctly to 29 of the 33 testing cases (87,9%), while the second one indicated correctly 31 of the same cases (93,9%). In the renal syndromes domain, however, the performance of HYCONES II - Simplified Fuzzy ARTMAP was superior to the one exhibited by CNM (p < 0,05). Both versions pointed out correctly, respectively, 108 (85%) and 95 (74.8%) diagnoses of the 127 testing cases presented to the system. HYCONES II - Simplified Fuzzy ARTMAP, therefore, displayed a satisfactory performance. However, the semantic contents of the neural nets it generated were completely different from the ones stemming from the CNM version. The networks that pointed out the final diagnosis in HYCONES - CNM were very similar to the knowledge graphs elicited from experts in congenital heart diseases. On the other hand, the networks activated in HYCONES II - Simplified Fuzzy ARTMAP operated with far more evidences than the CNM version. Besides this quantitative difference, there was a striking qualitative discrepancy among these two models. The Simplified Fuzzy ARTMAP version, even though pointing out to the correct diagnoses, used evidences that represented the complementary coding of the input pattern. This coding, inherent to the Simplified Fuzzy ARTMAP model, duplicates the input pattern, generating a new one depicting the evidence observed (on-cell) and, at the same time, the absent evidence, in relation to the total evidence employed to represent the input cases (off-cell). This coding shuts out the HYCONES explanation mechanism, since medical doctors usually reach a diagnostic conclusion rather from a set of observed evidences than from their absence. The next step taken was to improve the semantic contents of the Simplified Fuzzy ARTMAP model. To achieve this, the complement coding process was removed and the modified model was, then, revalidated, through the same testing sets as above described. In the CHD domain, the performance of HYCONES II - Simplified Fuzzy ARTMAP, without complementary coding, proved to be inferior to the one presented by CNM (p < 0,05). The first model singled out correctly 25 out of the 33 testing cases (75,8%), while the second one singled out correctly 31 out of the same 33 cases (93,9%). In the renal syndromes domain, the performances of HYCONES II - Simplified Fuzzy ARTMAP, without complementary coding, and HYCONES - CNM were similar. The first pointed out correctly to 98 of the 127 testing cases (77,2%), while the second one pointed out correctly to 95 of the same cases (74.8%). However, the recognition categories formed by this modified Simplified Fuzzy ARTMAP still presented quantitative and qualitative differences in their contents, when compared to the networks activated by CNM and to the knowledge graphs elicited from experts. This discrepancy, although smaller than the one observed in the original Fuzzy ARTMAP model, still restrained HYCONES explanation mechanism. The Semantic ART model (SMART) was, then, proposed. Its goal was to improve the semantic contents of ART recognition categories. To build this new model, the Simplified Fuzzy ARTMAP archictecture was preserved, while its learning algorithm was replaced by the CNM inductive learning mechanism (the punishments and rewards algorithm, associated with the pruning and normalization mechanisms). A new validation phase was, then, performed over the same testing sets. For the CHD domain, the perfomance comparison among SMART, Simplified Fuzzy ARTMAP, and CNM versions showed similar results. The first and the second versions pointed out correctly to 29 of the 33 testing cases (87,9%), while the third one singled out correctly 31 of the same testing cases (93,9%). For the renal syndromes domain, the performance of HYCONES II - SMART was superior to the one presented by the CNM version (p < 0,05), and equal to the performance presented by the Simplified Fuzzy ARTMAP version. SMART and Simplified Fuzzy ARTMAP singled out correctly 108 of the 127 testing cases (85%), while the CNM version pointed out correctly 95 of the same 127 testing cases (74.8%). Finally, it was observed that the neural networks generated by HYCONES II - SMART had a similar content to the networks generated by CNM and to the knowledge graphs elicited from multiple experts. The main contributions of this dissertation are: the design, implementation and validation of the Simplified Fuzzy ARTMAP and SMART models. The latter one, however, stands out for its learning mechanism, which provides a higher semantic value to the recognition categories, when compared to the categories formed by conventional ART models. This important enhancement is obtained by incorporating specificity and relevance concepts to ART's dynamics. This dissertation, however, represents not only the design and validation of two new connectionist models, but also, the enrichment of HYCONES. This is obtained through the continuation of a previous MSc dissertation, under the same supervision supervision. From the present work, therefore, it is given to the knowledge engineering, the choice among three different neural networks: CNM, Semantic ART and Simplified Fuzzy ARTMAP, all of which, display good performance. Indeed, the first and second models, in contrast to the third, support the context in a semantic way.

Les Fonctions biologiques sont le résultat de l'interaction de beaucoup de processus, avec differents objectives, complexités, niveaux d'hiérarchie, et changements de conditions que modi ent le comportement de systèmes. Nous utilisons des équations diferenciales ou dynamiques plus générales, et Stochastic Systèmes de Transition pour décrire la dynamique de changements des modèles. La composition, réconciliation et reutilisation des modèles nous permettent d'obtenir des descriptions de systèmes biologiques complètes et compatibles et leur combiner. Notre spéci cation de Systèmes Hybrides avec BioRica assures l'intégrité de modèles, et implement notre approche. Nous appliquons notre approche pour décrire in-silico deux systèmes: la dynamique de la fermentation du vin, et des décisions cellulaires associées à la formation de tissu d'os.

Large, software intensive systems are typically developed using a feature oriented development paradigm in which feature specifications are derived from domain requirements and features are implemented to satisfy such specifications. Historically, this approach has been followed in the telecommunications industry. More recently, in the automotive industry, features (for e.g. electronic stability control, collision avoidance etc.) are being developed as part of a software product line and a suitable subset of these features is integrated in an automobile model based on market requirements. Typically, features are designed independently by different engineering teams and are integrated later to create a system. Integrating features that are designed independently is extremely hard because the interactions between features are not understood properly and any incompatibilities may lead to costly redesign. In this thesis, we propose a framework for developing feature based systems such that even if features are incompatible, they can be integrated without redesign. Our view is that a feature based system consists of a base system and multiple features (or controllers), each of which independently advise the base system on how to react to an input so as to conform to their respective specifications. Such a system may reach a point of “conflict” between two or more features when they do not agree on a common action that the base system should perform. Instead of redesigning one or more features for resolving a conflict, we propose the novel notion of “conflicttolerance”, which requires features to be “resilient” or “tolerant” with regard to violations of their advice. Thus, unlike a classical feature, a conflicttolerant feature observes that its advice has been overridden, and takes this fact into account before proceeding to offer advice for subsequent behaviour of the base system. Conflict-tolerant features are composed using a priority order such that whenever a conflict occurs between two features, the base system continues with the advice of the higher priority feature. We guarantee that each feature is “maximally” utilized in that its advice is not taken only when there is a conflict with some higher priority controller. We show how to specify conflict-tolerant features for finite state, timed, and hybrid systems and also provide decision procedures for automated verification of finite state and timed systems. This provides a compositional technique for verifying systems which are composed of conflict-tolerant features. Our framework for developing feature based systems enables conflictresolution without redesign. The scope for reusing conflict tolerant features is significantly higher thus reducing design and verification effort.

El Raonament Basat en Casos (CBR) és un paradigma d'aprenentatge basat en establir analogies amb problemes prèviament resolts per resoldre'n de nous. Per tant, l'organització, l'accés i la utilització del coneixement previ són aspectes claus per tenir èxit en aquest procés. No obstant, la majoria dels problemes reals presenten grans volums de dades complexes, incertes i amb coneixement aproximat i, conseqüentment, el rendiment del CBR pot veure's minvat degut a la complexitat de gestionar aquest tipus de coneixement. Això ha fet que en els últims anys hagi sorgit una nova línia de recerca anomenada Soft-Computing and Intelligent Information Retrieval enfocada en mitigar aquests efectes. D'aquí neix el context d'aquesta tesi.Dins de l'ampli ventall de tècniques Soft-Computing per tractar coneixement complex, els Mapes Autoorganitzatius (SOM) destaquen sobre la resta per la seva capacitat en agrupar les dades en patrons, els quals permeten detectar relacions ocultes entre les dades. Aquesta capacitat ha estat explotada en treballs previs d'altres investigadors, on s'ha organitzat la memòria de casos del CBR amb SOM per tal de millorar la recuperació dels casos.La finalitat de la present tesi és donar un pas més enllà en la simple combinació del CBR i de SOM, de tal manera que aquí s'introdueixen les capacitats de Soft-Computing i de Knowledge Discovery de SOM en totes les fases del CBR per nodrir-les del nou coneixement descobert. A més a més, les mètriques de complexitat apareixen en aquest context com un instrument precís per modelar el funcionament de SOM segons la tipologia de les dades. L'assoliment d'aquesta integració es pot dividir principalment en quatre fites: (1) la definició d'una metodologia per determinar la millor manera de recuperar els casos tenint en compte la complexitat de les dades i els requeriments de l'usuari; (2) la millora de la fiabilitat de la proposta de solucions gràcies a les relacions entre els clústers i els casos; (3) la potenciació de les capacitats explicatives mitjançant la generació d'explicacions simbòliques; (4) el manteniment incremental i semi-supervisat de la memòria de casos organitzada per SOM.Tots aquests punts s'integren sota la plataforma SOMCBR, la qual és extensament avaluada sobre datasets provinents de l'UCI Repository i de dominis mèdics i telemàtics.Addicionalment, la tesi aborda de manera secundària dues línies de recerca fruït dels requeriments dels projectes on ha estat ubicada. D'una banda, s'aborda la definició de funcions de similitud específiques per definir com comparar un cas resolt amb un de nou mitjançant una variant de la Computació Evolutiva anomenada Evolució de Gramàtiques (GE). D'altra banda, s'estudia com definir esquemes de cooperació entre sistemes heterogenis per millorar la fiabilitat de la seva resposta conjunta mitjançant GE. Ambdues línies són integrades en dues plataformes, BRAIN i MGE respectivament, i són també avaluades amb els datasets anteriors. / El Razonamiento Basado en Casos (CBR) es un paradigma de aprendizaje basado en establecer analogías con problemas previamente resueltos para resolver otros nuevos. Por tanto, la organización, el acceso y la utilización del conocimiento previo son aspectos clave para tener éxito. No obstante, la mayoría de los problemas presentan grandes volúmenes de datos complejos, inciertos y con conocimiento aproximado y, por tanto, el rendimiento del CBR puede verse afectado debido a la complejidad de gestionarlos. Esto ha hecho que en los últimos años haya surgido una nueva línea de investigación llamada Soft-Computing and Intelligent Information Retrieval focalizada en mitigar estos efectos. Es aquí donde nace el contexto de esta tesis.Dentro del amplio abanico de técnicas Soft-Computing para tratar conocimiento complejo, los Mapas Autoorganizativos (SOM) destacan por encima del resto por su capacidad de agrupar los datos en patrones, los cuales permiten detectar relaciones ocultas entre los datos. Esta capacidad ha sido aprovechada en trabajos previos de otros investigadores, donde se ha organizado la memoria de casos del CBR con SOM para mejorar la recuperación de los casos.La finalidad de la presente tesis es dar un paso más en la simple combinación del CBR y de SOM, de tal manera que aquí se introducen las capacidades de Soft-Computing y de Knowledge Discovery de SOM en todas las fases del CBR para alimentarlas del conocimiento nuevo descubierto. Además, las métricas de complejidad aparecen en este contexto como un instrumento preciso para modelar el funcionamiento de SOM en función de la tipología de los datos. La consecución de esta integración se puede dividir principalmente en cuatro hitos: (1) la definición de una metodología para determinar la mejor manera de recuperar los casos teniendo en cuenta la complejidad de los datos y los requerimientos del usuario; (2) la mejora de la fiabilidad en la propuesta de soluciones gracias a las relaciones entre los clusters y los casos; (3) la potenciación de las capacidades explicativas mediante la generación de explicaciones simbólicas; (4) el mantenimiento incremental y semi-supervisado de la memoria de casos organizada por SOM. Todos estos puntos se integran en la plataforma SOMCBR, la cual es ampliamente evaluada sobre datasets procedentes del UCI Repository y de dominios médicos y telemáticos.Adicionalmente, la tesis aborda secundariamente dos líneas de investigación fruto de los requeri-mientos de los proyectos donde ha estado ubicada la tesis. Por un lado, se aborda la definición de funciones de similitud específicas para definir como comparar un caso resuelto con otro nuevo mediante una variante de la Computación Evolutiva denominada Evolución de Gramáticas (GE). Por otro lado, se estudia como definir esquemas de cooperación entre sistemas heterogéneos para mejorar la fiabilidad de su respuesta conjunta mediante GE. Ambas líneas son integradas en dos plataformas, BRAIN y MGE, las cuales también son evaluadas sobre los datasets anteriores. / Case-Based Reasoning (CBR) is an approach of machine learning based on solving new problems by identifying analogies with other previous solved problems. Thus, organization, access and management of this knowledge are crucial issues for achieving successful results. Nevertheless, the major part of real problems presents a huge amount of complex data, which also presents uncertain and partial knowledge. Therefore, CBR performance is influenced by the complex management of this knowledge. For this reason, a new research topic has appeared in the last years for tackling this problem: Soft-Computing and Intelligent Information Retrieval. This is the point where this thesis was born.Inside the wide variety of Soft-Computing techniques for managing complex data, the Self-Organizing Maps (SOM) highlight from the rest due to their capability for grouping data according to certain patterns using the relations hidden in data. This capability has been used in a wide range of works, where the CBR case memory has been organized with SOM for improving the case retrieval.The goal of this thesis is to take a step up in the simple combination of CBR and SOM. This thesis presents how to introduce the Soft-Computing and Knowledge Discovery capabilities of SOM inside all the steps of CBR to promote them with the discovered knowledge. Furthermore, complexity measures appear in this context as a mechanism to model the performance of SOM according to data topology. The achievement of this goal can be split in the next four points: (1) the definition of a methodology for setting up the best way of retrieving cases taking into account the data complexity and user requirements; (2) the improvement of the classification reliability through the relations between cases and clusters; (3) the promotion of the explaining capabilities by means of the generation of symbolic explanations; (4) the incremental and semi-supervised case-based maintenance. All these points are integrated in the SOMCBR framework, which has been widely tested in datasets from UCI Repository and from medical and telematic domains. Additionally, this thesis secondly tackles two additional research lines due to the requirements of a project in which it has been developed. First, the definition of similarity functions ad hoc a domain is analyzed using a variant of the Evolutionary Computation called Grammar Evolution (GE). Second, the definition of cooperation schemes between heterogeneous systems is also analyzed for improving the reliability from the point of view of GE. Both lines are developed in two frameworks, BRAIN and MGE respectively, which are also evaluated over the last explained datasets.

Solid oxide fuel cell/gas turbine (SOFC/GT) hybrid systems harness the capability to operate nearly 15 to 20 percentage points more efficiently than standard natural gas or pulverized coal power plants. Though the performance of these systems is quite promising, a number of system integration challenges, primarily with regards to thermal transport, still remain. It is for that reason that the Hybrid Performance Project (HyPer) facility, a Hardware-in-the-Loop SOFC/GT hybrid simulator, was built at the National Energy Technology Laboratory in Morgantown, WV. The HyPer facility couples an actual gas turbine with a combination of hardware and software that are used to simulate an actual SOFC. The facility is used to empirically address the system integration issues associated with fuel cell/gas turbine hybrids. Through this dissertation project, the software component of the SOFC simulator was upgraded from a 0-D lumped SOFC model to a 1-D, distributed, real-time operating SOFC model capable of spatio-temporal characterization of a fuel cell operating with a gas turbine in a hybrid arrangement. Once completed and verified, the upgraded HyPer facility was used to characterize the impact of cold air by-pass and initial fuel cell load on electrochemical start-up in an SOFC/GT hybrid environment. The impact of start-up on fuel cell inlet process parameters, SOFC performance and SOFC distributed behavior are presented and analyzed in comparative manner. This study represents the first time that an empirical parametric study, characterizing system operation during electrochemical start-up has been conducted.

In this dissertation, we present several main research thrusts involving thermodynamic stabilization via energy dissipating hybrid controllers and nonlinear control of network systems. Specifically, a novel class of fixed-order, energy-based hybrid controllers is presented as a means for achieving enhanced energy dissipation in Euler-Lagrange, lossless, and dissipative dynamical systems. These dynamic controllers combine a logical switching architecture with continuous dynamics to guarantee that the system plant energy is strictly decreasing across switching. In addition, we construct hybrid dynamic controllers that guarantee that the closed-loop system is consistent with basic thermodynamic principles. In particular, the existence of an entropy function for the closed-loop system is established that satisfies a hybrid Clausius-type inequality. Special cases of energy-based hybrid controllers involving state-dependent switching are described, and the framework is applied to aerospace system models. The overall framework demonstrates that energy-based hybrid resetting controllers provide an extremely efficient mechanism for dissipating energy in nonlinear dynamical systems. Next, we present finite-time coordination controllers for multiagent network systems. Recent technological advances in communications and computation have spurred a broad interest in autonomous, adaptable vehicle formations. Distributed decision-making for coordination of networks of dynamic agents addresses a broad area of applications including cooperative control of unmanned air vehicles, microsatellite clusters, mobile robotics, and congestion control in communication networks. In this part of the dissertation we focus on finite-time consensus protocols for networks of dynamic agents with undirected information flow. The proposed controller architectures are predicated on the recently developed notion of system thermodynamics resulting in thermodynamically consistent continuous controller architectures involving the exchange of information between agents that guarantee that the closed-loop dynamical network is consistent with basic thermodynamic principles.

Η παρούσα διπλωματική εργασία έχει ως αντικείμενο την μελέτη και τη μοντελοποίηση ενός υβριδικού συστήματος, που αποτελείται από φωτοβολταϊκά και κυψέλες καυσίμου τα οποία τροφοδοτούν ένα μικροδίκτυο. Το φορτίο το οποίο καλείται το σύστημα να καλύψει είναι 30kW και 15kVar. Για τη μοντελοποίηση του συστήματος χρησιμοποιείται το πρόγραμμα σχεδίασης και προσομοίωσης ηλεκτρικών συστημάτων PSCAD. Στην εργασία αυτή το ενδιαφέρον επικεντρώνεται στη συμπεριφορά του συστήματος στη μόνιμη κατάσταση λειτουργίας και στη συμπεριφορά σε κάποια μεταβατατικά φαινόμενα. Τα μεταβατικά φαινόμενα που εξετάζονται είναι τα εξής: - Συμπεριφορά συστήματος σε μεταβολές της ηλιοφάνειας. - Απόκριση συστήματος σε μεταβολές του φορτίου. - Τριφασικό και μονοφασικό βραχυκύκλωμα με τη γη στη γραμμή. Στο κεφάλαιο 1 γίνεται αναφορά στις ανανεώσιμες πηγές, στην εξέλιξή και στην προοπτική τους στο μέλλον. Στη συνέχεια περιγράφονται τα διάφορα είδη των υβριδικών συστημάτων και γίνεται αναφορά για τη διεσπαρμένη παραγωγή και τα μικροδίκτυα. Στο κεφάλαιο 2 περιγράφεται η λειτουργία των φωτοβολταϊκών κυττάρων, παρουσιάζονται οι I-V και P-V χαρακτηριστικές και στη συνέχεια ακολουθεί αναλυτική περιγραφή του μοντέλου της φωτοβολταϊκής συστοιχίας που θα χρησιμοποιήσουμε στην προσομοίωση στο PScad. Στο κεφάλαιο 3 περιγράφεται η λειτουργία των κυψελών καυσίμων και παρουσιάζονται τα διάφορα είδη κυψελών που υπάρχουν στις μέρες μας. Εν συνεχεία ακολουθεί αναλυτική περιγραφή της λειτουργίας της κυψέλης καυσίμου πολυμερούς ηλεκτρολυτικής μεμβράνης PEM FC καθώς αυτό το είδος των κυψελών καυσίμου χρησιμοποιούμε στο σύστημά μας. Τέλος, κατασκευάζεται το μοντέλο της PEM FC και ακολουθεί η αναλυτική περιγραφή του μοντέλου. Στο κεφάλαιο 4 γίνεται αναφορά στους dc/dc μετατροπείς ανύψωσης τάσης και στις περιοχές λειτουργίας τους. Εν συνεχεία γίνεται διαστασολόγηση του ανυψωτή τάσης και παρουσιάζεται το σύστημα ελέγχου του. Επίσης μελετάται η λειτουργία του ανυψωτή ως ανιχνευτή του σημείου μέγιστης ισχύος της φωτοβολταϊκής συστοιχίας και γίνεται προσομοίωση των κυψελών και της φωτοβολταϊκής συστοιχίας με τους dc/dc μετατροπείς που χρησιμοποιούνται. Στο κεφάλαιο 5 ακολουθεί η περιγραφή της λειτουργίας των μονοφασικών και τριφασικών αντιστροφέων DC/AC και παρουσιάζεται αναλυτικά η στρατηγική της διαμόρφωσης εύρους παλμών (PWM). Τέλος, γίνεται υπολογισμός των τιμών του φίλτρου που τοποθετείται στην έξοδο του αντιστροφέα και στη συνέχεια προσομοιώνεται και μελετάται το μοντέλο του αντιστροφέα που χρησιμοποιούμε στο σύστημά μας. Στο κεφάλαιο 6 γίνεται η πλήρη προσομοίωση του συστήματος. Περιγράφονται επίσης και τα μοντέλα του φορτίου, του Μ/Σ, του δικτύου και της μηχανής παραγωγής σφαλμάτων. Τέλος, παρουσιάζονται και περιγράφονται τα αποτελέσματα των μελετών που πραγματοποιούνται στο σύστημα (απόκριση συστήματος σε κανονικές συνθήκες, απόκριση συστήματος για μεταβολή της ηλιοφάνειας, μελέτη συστήματος στη μεταβολή του φορτίου, προσομοίωση συστήματος για σφάλματα στη γραμμή). / The current diploma thesis presents the study and simulation of a hybrid system, which consists of a Photovoltaic Array (PV) and a Fuel Cell stack (FC), which supports a microgrid. The critical load that the system supports is 30kW and 15kVar. For the system modelling the program of designing and simulation of electric systems PSCAD is used. At this project the interest is focused in the behavior of system in the permanent situation of operation and in the behavior in certain transient phenomena. The transient phenomena that are examined are the following: - Behavior of the system in variations in solar insolation. - System’s response to variations of the load. - Fault analysis (single line to ground and three-phase line to ground short circuit). In Chapter 1 there is a description of renewable energy sources, their development and prospects in the future. Afterwards, the different kinds of hybrid systems are mentioned and there is a reference in distributed generation and microgrids. In Chapter 2 the function of photovoltaic cells and different kinds of photovoltaic technologies are described. Furthermore, I-V and P-V characteristic are presented and then follows an analytical description of the photovoltaic model that we use in our simulation in PScad. In Chapter 3 , the function of the Fuel Cells is described and the different kinds of fuel cells are presented. Afterwards, there is an analytical description of the proton exchange membrane fuel cells (PEM FC) as we use this kind of FC in our system. Finally, the PEM FC model is constructed and an analytical description of the model is made. The Chapter 4 entails information for the dc/dc boost converters and their operating modes. Subsequently the control system of the dc/dc boost converter is presented. Finally, the utilization of the boost converter as a maximum power point tracker is examined and the simulation of the FV and PV with the boost converter is presented. In Chapter 5 there is a description of the function of the single-phase and three-phase dc/ac inverters and the technique of the pulse-width modulation (PWM) is presented. Finally, the value of the output filter of the inverter is calculated and then the model of the dc/ac inverter, which we use in our system, is simulated. In Chapter 6 there is a simulation of the whole system. Furthermore, the models of the load, the transformer and the fault generator are described. Finally, the results of the simulations for which the system was tested is presented (system response to normal conditions, system response to changes in the insolation, System’s response to variations of the load, system response to grid faults).

Les travaux présentés dans cette thèse concernent le problème d'identification des systèmes à retards et d'une certaine classe de systèmes hybrides appelés systèmes "impulsifs".Dans la première partie, un algorithme d'identification rapide a été proposé pour les systèmes à entrée retardée. Il est basé sur une méthode d'estimation distributionnelle non asymptotique initiée pour les systèmes sans retard. Une telle technique mène à des schémas de réalisation simples, impliquant des intégrateurs, des multiplicateurs et des fonctions continues par morceaux polynomiales ou exponentielles. Dans le but de généraliser cette approche pour les systèmes à retard, trois exemples d'applications ont été étudiées. La deuxième partie a été consacrée à l'identification des systèmes impulsifs. En se basant sur le formalisme des distributions, une procédure d'identification a été élaborée afin d'annihiler les termes singuliers des équations différentielles représentant ces systèmes. Par conséquent, une estimation en ligne des instants de commutations et des paramètres inconnus est prévue indépendamment des lois de commutations. Des simulations numériques d'un pendule simple soumis à des frottements secs illustrent notre méthodologie / This PhD thesis concerns the problem of identification of the delays systems and the continuous-time systems subject to impulsive terms.Firstly, a fast identification algorithm is proposed for systems with delayed inputs. It is based on a non-asymptotic distributional estimation technique initiated in the framework of systems without delay. Such technique leads to simple realization schemes, involving integrators, multipliers andContribution to the identification of time delays systems and a class of hybrid systems piecewise polynomial or exponential time functions. Thus, it allows for a real time implementation. In order to introduce a generalization to systems with input delay, three simple examples are presented.The second part deals with on-line identification of continuous-time systems subject to impulsive terms. Using a distribution framework, a scheme is proposed in order to annihilate singular terms in differential equations representing a class of impulsive systems. As a result, an online estimation of unknown parameters is provided, regardless of the switching times or the impulse rules. Numerical simulations of simple pendulum subjected to dry friction are illustrating our methodology

L’objectif de cette thèse est de dimensionner et régler un système hybride de production d’énergie pour un site isolé de type insulaire, basé sur des énergies renouvelables marines. De manière préliminaire divers systèmes de production d’énergie renouvelable marine ont d’abord été étudiés et comparés de manière qualitative à des systèmes de production d’énergie classiques. Plusieurs types de système de stockage d’énergie ont également été étudiés, comparés et évalués dans le cas du site considéré. Cette analyse préliminaire a été étendue aux différents types de transmissions d’énergie offshore et de méthodes de réglage des convertisseurs associés aux sources renouvelables. A partir de l’étude des caractéristiques du site et de l’analyse statistique des ressources renouvelables (vents, courants marins) une méthode de dimensionnement des éléments du système de production est présentée, dans l’objectif de minimiser les émissions de CO2 et le coût du système sur son cycle de vie. Pour cela, une solution de gestion de la puissance basée sur la logique floue est proposée pour le type de site considéré et comparée à une solution plus classique basée sur des règles logiques. Pour finir, une étude détaillée des différentes méthodes de réglage du système hybride côté réseau est présentée. Trois niveaux de réglage sont considérés : réglage d’une source unique, réglage d’une ferme de plusieurs sources et réglage global du système hybride. Plusieurs modes de réglage sont considérés pour chaque niveau. / This PhD thesis models a renewable-based hybrid power supply system applied in an islanded context and investigates sizing and regulation strategies of such a hybrid system. First, various marine energy production technologies were reviewed and compared to common renewable resources. As well, various energy storage technologies were reviewed, compared, and evaluated to fit the chosen site characteristics. A brief investigation on offshore energy transmission and inverter regulations methods is presented. Then, a study of the site characteristics, and the availability of the different renewable energy resources in the area are presented. This energy study constitutes the basis of the proposed system sizing method, where minimizing the cost and the CO2 emissions are considered as the main objectives. Furthermore, a fuzzy logic power management approach is proposed for the islanded microgrid. Finally, a detailed study of the system components grid-side inverter regulation is presented. Three regulation levels were investigated: the single inverter, the renewable farm, and the hybrid system. In this context, different regulation strategies are considered at each level.