Individualmente, as linhas de pesquisa da Inteligência Artificial têm proposto abordagens para a resolução de inúmeros problemas complexos do mundo real. O paradigma orientado a agentes provê os agentes autônomos, capazes de perceber os seus ambientes, reagir de acordo com diferentes circunstâncias e estabelecer interações sociais com outros agentes de software ou humanos. As redes Bayesianas fornecem uma maneira de representar graficamente as distribuições de probabilidades condicionais e permitem a realização de raciocínios probabilísticos baseados em evidências. As ontologias são especificações explícitas e formais de conceituações, que são usadas em uma variedade de áreas de pesquisa, incluindo os Sistemas Multiagentes. Contudo, existem aplicações cujos requisitos não podem ser atendidos por uma única tecnologia. Circunstâncias como estas exigem a integração de tecnologias desenvolvidas por distintas áreas da Ciência da Computação. Esta dissertação trata a integração do modelo de agentes BDI (Belief-Desire-Intention) e das redes Bayesianas. Além disso, é adotada uma abordagem baseada em ontologias para representar o conhecimento incerto dos agentes. O primeiro passo em direção a integração foi o desenvolvimento de uma ontologia para representar a estrutura das redes Bayesinas. Esta ontologia tem como principal objetivo permitir a interoperabilidade agentes compatíveis com a arquitetura proposta. No entanto, a ontologia também facilita o entendimento necessário para abstrair os estados mentais e processos cognitivos dos agentes através de elementos das redes Bayesianas. Uma vez construída a ontologia, a mesma foi integrada com a arquitetura BDI. Através da integração do modelo BDI com as redes Bayesianas foi obtida uma arquitetura cognitiva de agentes capaz de deliberar sob incerteza. O processo de integração foi composto de duas etapas: abstração dos estados mentais através de elementos das redes Bayesianas e especificação do processo deliberativo. Finalmente, foi desenvolvido um estudo de caso, que consistiu na aplicação da arquitetura proposta no Agente Social, um componente de um portal educacional multiagente (PortEdu). / Individually, Artificial Intelligence research areas have proposed approaches to solve several complex real-world problems. The agent-based paradigm provided autonomous agents, capable of perceiving their environment, reacting in accordance with different situations, and establishing social interactions with other software agents and humans. Bayesian networks provided a way to represent graphically the conditional probability distributions and an evidence-based probabilistic reasoning. Ontologies are an effort to develop formal and explicit specifications of concepts, which have been used by a wide range of research areas, including Multiagent Systems. However, there are applications whose requirements can not be addressed by a single technology. Circumstances like these demand the integration of technologies developed by distinct areas of Computer Science. This work is particularly concerned with the integration of Belief-Desire-Intention (BDI) agent architecture and Bayesian networks. Moreover, it is adopted an ontology-based approach to represent the agent’s uncertain knowledge. To bring together those technologies, it was developed an ontology to represent the structure of Bayesian networks knowledge representation. This ontology supports the interoperability among agents that comply with the proposed architecture, and it also facilitates the understanding necessary to abstract the agents’ mental states and cognitive processes through elements of Bayesian networks. Once specified the ontology, it was integrated with the BDI agent architecture. By integrating BDI architecture and Bayesian networks, it was obtained a cognitive agent architecture capable of reasoning under uncertainty. It was performed in two stages: abstraction of mental states through Bayesian networks and specification of the deliberative process. Finally, it was developed a case study, which consists in applying the probabilistic BDI architecture in the Social Agent, a component of a multiagent educational portal (PortEdu).

Individualmente, as linhas de pesquisa da Inteligência Artificial têm proposto abordagens para a resolução de inúmeros problemas complexos do mundo real. O paradigma orientado a agentes provê os agentes autônomos, capazes de perceber os seus ambientes, reagir de acordo com diferentes circunstâncias e estabelecer interações sociais com outros agentes de software ou humanos. As redes Bayesianas fornecem uma maneira de representar graficamente as distribuições de probabilidades condicionais e permitem a realização de raciocínios probabilísticos baseados em evidências. As ontologias são especificações explícitas e formais de conceituações, que são usadas em uma variedade de áreas de pesquisa, incluindo os Sistemas Multiagentes. Contudo, existem aplicações cujos requisitos não podem ser atendidos por uma única tecnologia. Circunstâncias como estas exigem a integração de tecnologias desenvolvidas por distintas áreas da Ciência da Computação. Esta dissertação trata a integração do modelo de agentes BDI (Belief-Desire-Intention) e das redes Bayesianas. Além disso, é adotada uma abordagem baseada em ontologias para representar o conhecimento incerto dos agentes. O primeiro passo em direção a integração foi o desenvolvimento de uma ontologia para representar a estrutura das redes Bayesinas. Esta ontologia tem como principal objetivo permitir a interoperabilidade agentes compatíveis com a arquitetura proposta. No entanto, a ontologia também facilita o entendimento necessário para abstrair os estados mentais e processos cognitivos dos agentes através de elementos das redes Bayesianas. Uma vez construída a ontologia, a mesma foi integrada com a arquitetura BDI. Através da integração do modelo BDI com as redes Bayesianas foi obtida uma arquitetura cognitiva de agentes capaz de deliberar sob incerteza. O processo de integração foi composto de duas etapas: abstração dos estados mentais através de elementos das redes Bayesianas e especificação do processo deliberativo. Finalmente, foi desenvolvido um estudo de caso, que consistiu na aplicação da arquitetura proposta no Agente Social, um componente de um portal educacional multiagente (PortEdu). / Individually, Artificial Intelligence research areas have proposed approaches to solve several complex real-world problems. The agent-based paradigm provided autonomous agents, capable of perceiving their environment, reacting in accordance with different situations, and establishing social interactions with other software agents and humans. Bayesian networks provided a way to represent graphically the conditional probability distributions and an evidence-based probabilistic reasoning. Ontologies are an effort to develop formal and explicit specifications of concepts, which have been used by a wide range of research areas, including Multiagent Systems. However, there are applications whose requirements can not be addressed by a single technology. Circumstances like these demand the integration of technologies developed by distinct areas of Computer Science. This work is particularly concerned with the integration of Belief-Desire-Intention (BDI) agent architecture and Bayesian networks. Moreover, it is adopted an ontology-based approach to represent the agent’s uncertain knowledge. To bring together those technologies, it was developed an ontology to represent the structure of Bayesian networks knowledge representation. This ontology supports the interoperability among agents that comply with the proposed architecture, and it also facilitates the understanding necessary to abstract the agents’ mental states and cognitive processes through elements of Bayesian networks. Once specified the ontology, it was integrated with the BDI agent architecture. By integrating BDI architecture and Bayesian networks, it was obtained a cognitive agent architecture capable of reasoning under uncertainty. It was performed in two stages: abstraction of mental states through Bayesian networks and specification of the deliberative process. Finally, it was developed a case study, which consists in applying the probabilistic BDI architecture in the Social Agent, a component of a multiagent educational portal (PortEdu).

Individualmente, as linhas de pesquisa da Inteligência Artificial têm proposto abordagens para a resolução de inúmeros problemas complexos do mundo real. O paradigma orientado a agentes provê os agentes autônomos, capazes de perceber os seus ambientes, reagir de acordo com diferentes circunstâncias e estabelecer interações sociais com outros agentes de software ou humanos. As redes Bayesianas fornecem uma maneira de representar graficamente as distribuições de probabilidades condicionais e permitem a realização de raciocínios probabilísticos baseados em evidências. As ontologias são especificações explícitas e formais de conceituações, que são usadas em uma variedade de áreas de pesquisa, incluindo os Sistemas Multiagentes. Contudo, existem aplicações cujos requisitos não podem ser atendidos por uma única tecnologia. Circunstâncias como estas exigem a integração de tecnologias desenvolvidas por distintas áreas da Ciência da Computação. Esta dissertação trata a integração do modelo de agentes BDI (Belief-Desire-Intention) e das redes Bayesianas. Além disso, é adotada uma abordagem baseada em ontologias para representar o conhecimento incerto dos agentes. O primeiro passo em direção a integração foi o desenvolvimento de uma ontologia para representar a estrutura das redes Bayesinas. Esta ontologia tem como principal objetivo permitir a interoperabilidade agentes compatíveis com a arquitetura proposta. No entanto, a ontologia também facilita o entendimento necessário para abstrair os estados mentais e processos cognitivos dos agentes através de elementos das redes Bayesianas. Uma vez construída a ontologia, a mesma foi integrada com a arquitetura BDI. Através da integração do modelo BDI com as redes Bayesianas foi obtida uma arquitetura cognitiva de agentes capaz de deliberar sob incerteza. O processo de integração foi composto de duas etapas: abstração dos estados mentais através de elementos das redes Bayesianas e especificação do processo deliberativo. Finalmente, foi desenvolvido um estudo de caso, que consistiu na aplicação da arquitetura proposta no Agente Social, um componente de um portal educacional multiagente (PortEdu). / Individually, Artificial Intelligence research areas have proposed approaches to solve several complex real-world problems. The agent-based paradigm provided autonomous agents, capable of perceiving their environment, reacting in accordance with different situations, and establishing social interactions with other software agents and humans. Bayesian networks provided a way to represent graphically the conditional probability distributions and an evidence-based probabilistic reasoning. Ontologies are an effort to develop formal and explicit specifications of concepts, which have been used by a wide range of research areas, including Multiagent Systems. However, there are applications whose requirements can not be addressed by a single technology. Circumstances like these demand the integration of technologies developed by distinct areas of Computer Science. This work is particularly concerned with the integration of Belief-Desire-Intention (BDI) agent architecture and Bayesian networks. Moreover, it is adopted an ontology-based approach to represent the agent’s uncertain knowledge. To bring together those technologies, it was developed an ontology to represent the structure of Bayesian networks knowledge representation. This ontology supports the interoperability among agents that comply with the proposed architecture, and it also facilitates the understanding necessary to abstract the agents’ mental states and cognitive processes through elements of Bayesian networks. Once specified the ontology, it was integrated with the BDI agent architecture. By integrating BDI architecture and Bayesian networks, it was obtained a cognitive agent architecture capable of reasoning under uncertainty. It was performed in two stages: abstraction of mental states through Bayesian networks and specification of the deliberative process. Finally, it was developed a case study, which consists in applying the probabilistic BDI architecture in the Social Agent, a component of a multiagent educational portal (PortEdu).

Introducción: La depresión es un problema de salud pública, puesto que representa la primera causa de enfermedad psiquiátrica discapacitante en el Perú. Asimismo, la presencia de síntomas depresivos podría modificar la adherencia al tratamiento y mayor probabilidad de mal control glicémico en pacientes con Diabetes Mellitus tipo 2 (DM2). Materiales y Métodos: Se realizó un estudio trasversal, analítico. Se halló un tamaño muestral de 98 pacientes con diagnóstico de DM2 que acudieron al consultorio de endocrinología de un hospital del seguro social en el periodo de Octubre - Noviembre del año 2014. Se utilizaron tres escalas, la principal fue la del Inventario de Depresión de Beck II como variable de respuesta; las otras dos escalas, Escala de Áreas Problemáticas en la Diabetes (PAID) y la escala de Morisky-Green, fueron utilizadas como respuesta complementaria. En el estudio se definió como buen control glicémico una hemoglobina glicosilada ≤7. Resultados: Se estudió 104 pacientes, predominando el sexo femenino (53,8%), con edad promedio de 56 años con tendencia a ser menores de 60 años (p>0,05), casados y de grado superior. El índice de masa corporal tuvo asociación con la presencia de síntomas depresivos (p 0,003), y se encontró mayor frecuencia en aquellos con obesidad de tipo II y III. El 57,7% de la población tuvo mal control glicémico, de los cuales el 73,3% presentaban síntomas depresivos (p<0,05), y el 84,1% no tuvo adherencia al tratamiento (p<0,05). Se encontró una asociación significativa entre el mal control glicémico y la presencia de síntomas depresivos independientemente de variables potencialmente confusoras con una razón de prevalencias de 1,9 (1,1 a 3,6). Conclusiones: Se concluye la existencia de asociación entre el mal control glicémico y la presencia de síntomas depresivos en pacientes con DM2. De tal forma, se sugiere brindar un enfoque integral en dichos pacientes, en los cuales también deberá evaluarse el factor psicológico. / Introduction: Depression is an important problem of public health because it is the primary cause of disabling psychiatric illness in Peru. Also, the presence of depressive symptoms could modify the adherence to the treatment and a high probability of a bad glycemic control in patients with Diabetes Mellitus type 2 (DM2). Methods: A cross-sectional analytical study was done, with 98 patients who were diagnosed with DM2, who went to the Endocrinology Department during the period of October-November 2014 in a social security hospital. We used three scales, the first one was The Beck Depression Inventory II which evaluated the presence of depressive symptoms; the scale of Problem Areas in Diabetes (PAID) evaluates the emotional stress they faced during the illness, and the Morisky-Green Scale is the third scale used which valued the adherence to the pharmacological treatment. Results: in population, the predominance was on female sex (53,8%), with a mean age of 56 years tended to be younger than 60 years (p> 0.05), married and a superior grade. BMI was associated with the presence of depressive symptoms (p 0.003), and most often found in those with obesity, type II and III. The population had 57.7% of poor glycemic control, of which 73.3% had depressive symptoms (p <0.05), and 84.1% had no adherence to treatment (p <0.05). It was found a significant association between bad glycemic control and presence of depressive symptoms independently of the potentially confounding variables with a prevalence ratio to 1.9 (1, 1 -3, 6). Conclusions: It is concluded the association between the bad glycemic control and the presence of depressive symptoms in diabetic patients. Therefore, it is suggested to use an integral approach on such patients who will be subject to a psychological evaluation.

Cyber-physical systems refer to a new generation of synergy systems with integrated computational and physical processes which interact with one other. The development and simulation of cyber-physical systems (CPSs) are obstructed by the complexity of the subsystems of which they are comprised, fundamental differences in the operation of cyber and physical elements, significant correlative dependencies among the elements, and operation in dynamic and open environments. The Multiple Belief-Desire-Intention (BDI) agent system (BDI multi-agent system) is a promising choice for overcoming these challenges, since it offers a natural way to decompose complex systems or large scale problems into decentralized, autonomous, interacting, more or less intelligent entities. In particular, BDI agents have the ability to interact with, and expand the capabilities of, the physical world through computation, communication, and control. A BDI agent has its philosophical grounds on intentionality and practical reasoning, and it is natural to combine a philosophical model of human practical reasoning with the physical operation and any cyber infrastructure. In this thesis, we introduce the BDI Model, discuss implementations of BDI agents from an ideal theoretical perspective as well as from a more practical perspective, and show how they can be used to bridge the cyber infrastructure and the physical operation using the framework. We then strengthen the framework's performance using the state-of-the-art parallel computing architecture and eventually propose a BDI agent based software framework to enable the efficient modeling and simulation of heterogeneous CPS systems in an integrated manner. / Electrical and Computer Engineering

Sequent calculi for BDI logics is a research object of the thesis. BDI logics are widely used for agent system description and implementation. Agents are autonomous systems, those acts in some environment and aspire to achieve preassigned goals. Implementation of the decision making is the main and the most complicated part in agent systems implementation. Logic calculi may be used for the decision making implementation. In this thesis, there are researched sequent calculi for BDI logics. Sequent calculi for BDI logics, like sequent calculi for other modal logics, use loop-check technique to get decidability. Inefficient loop-check takes a major part of the resources used for the derivation. For some modal logics, there are known loop-check free sequent calculi or calculi with an efficient loop-check. In this thesis, there is presented loop-check free sequent calculus for KD45 logic, which is the main fragment of the BDI logics. Introduced calculus not only eliminates loop-check, but also simplifies sequent derivation. For the branching time logic (another BDI logic fragment) there is presented sequent calculus with an efficient loop-check. Obtained results are adapted for creation sequent calculi for monoagent and multiagent BDI logics. Introduced calculi use only restricted loop-check. Moreover, loop-check is totally eliminated for some types of the loops. These results enables to create more efficient agent systems, those are based on the BDI logics. / Darbe nagrinėjami sekvenciniai skaičiavimai BDI logikoms. BDI logikos yra plačiai naudojamos agentinių sistemų aprašymui ir realizavimui. Agentai yra autonomiškos sistemos, kurios veikia kažkurioje aplinkoje ir siekia įvykdyti iš anksto apibrėžtus tikslus. Sprendimų priėmimo realizavimas yra svarbiausia ir sudėtingiausia dalis realizuojant agentines sistemas. Sprendimo priėmimo realizavimui gali būti naudojami logikos skaičiavimai. Šiame darbe ir yra nagrinėjami sekvenciniai skaičiavimai BDI logikoms. BDI logikose, kaip ir kitose modalumo logikose, yra naudojama ciklų paieška išsprendžiamumui gauti. Neefektyvi ciklų paieška užima didesnę išvedimų paieškos resursų dalį. Kai kurioms modalumo logikoms yra žinomi becikliai skaičiavimai ar skaičiavimai naudojantys efektyvią ciklų paiešką. Šiame darbe yra pateikiamas beciklis sekvencinis skaičiavimas KD45 logikai, kuri yra esminis BDI logikų fragmentas. Pateiktas skaičiavimas ne tik eliminuoja ciklų paiešką, bet ir supaprastina patį sekvencijos išvedimą. Skaidaus laiko logikai (kitam BDI logikų fragmentui) yra pateikiamas sekvencinis skaičiavimas naudojantis efektyvią ciklų paiešką. Gauti rezultatai yra pritaikyti sukuriant sekvencinius skaičiavimus vianaagentinei ir daugiaagentinei BDI logikoms. Pristatyti skaičiavimai naudoja tik apribotą ciklų paiešką. Be to, kai kurių tipų ciklus eliminuoja visiškai. Šie rezultatai įgalina kurti efektyvesnes agentines sistemas, paremtas BDI logikomis.

Darbe nagrinėjami sekvenciniai skaičiavimai BDI logikoms. BDI logikos yra plačiai naudojamos agentinių sistemų aprašymui ir realizavimui. Agentai yra autonomiškos sistemos, kurios veikia kažkurioje aplinkoje ir siekia įvykdyti iš anksto apibrėžtus tikslus. Sprendimų priėmimo realizavimas yra svarbiausia ir sudėtingiausia dalis realizuojant agentines sistemas. Sprendimo priėmimo realizavimui gali būti naudojami logikos skaičiavimai. Šiame darbe ir yra nagrinėjami sekvenciniai skaičiavimai BDI logikoms. BDI logikose, kaip ir kitose modalumo logikose, yra naudojama ciklų paieška išsprendžiamumui gauti. Neefektyvi ciklų paieška užima didesnę išvedimų paieškos resursų dalį. Kai kurioms modalumo logikoms yra žinomi becikliai skaičiavimai ar skaičiavimai naudojantys efektyvią ciklų paiešką. Šiame darbe yra pateikiamas beciklis sekvencinis skaičiavimas KD45 logikai, kuri yra esminis BDI logikų fragmentas. Pateiktas skaičiavimas ne tik eliminuoja ciklų paiešką, bet ir supaprastina patį sekvencijos išvedimą. Skaidaus laiko logikai (kitam BDI logikų fragmentui) yra pateikiamas sekvencinis skaičiavimas naudojantis efektyvią ciklų paiešką. Gauti rezultatai yra pritaikyti sukuriant sekvencinius skaičiavimus vianaagentinei ir daugiaagentinei BDI logikoms. Pristatyti skaičiavimai naudoja tik apribotą ciklų paiešką. Be to, kai kurių tipų ciklus eliminuoja visiškai. Šie rezultatai įgalina kurti efektyvesnes agentines sistemas, paremtas BDI logikomis. / Sequent calculi for BDI logics is a research object of the thesis. BDI logics are widely used for agent system description and implementation. Agents are autonomous systems, those acts in some environment and aspire to achieve preassigned goals. Implementation of the decision making is the main and the most complicated part in agent systems implementation. Logic calculi may be used for the decision making implementation. In this thesis, there are researched sequent calculi for BDI logics. Sequent calculi for BDI logics, like sequent calculi for other modal logics, use loop-check technique to get decidability. Inefficient loop-check takes a major part of the resources used for the derivation. For some modal logics, there are known loop-check free sequent calculi or calculi with an efficient loop-check. In this thesis, there is presented loop-check free sequent calculus for KD45 logic, which is the main fragment of the BDI logics. Introduced calculus not only eliminates loop-check, but also simplifies sequent derivation. For the branching time logic (another BDI logic fragment) there is presented sequent calculus with an efficient loop-check. Obtained results are adapted for creation sequent calculi for monoagent and multiagent BDI logics. Introduced calculi use only restricted loop-check. Moreover, loop-check is totally eliminated for some types of the loops. These results enables to create more efficient agent systems, those are based on the BDI logics.

In this thesis we extend BDI logics, which are normal multimodal logics with an arbitrary set of normal modal operators, from three different perspectives. Firstly, based on some recent developments in modal logic, we examine BDI logics from a combining logic perspective and apply combination techniques like fibring/dovetailing for explaining them. The second perspective is to extend the underlying logics so as to include action constructs in an explicit way based on some recent action-related theories. The third perspective is to adopt a non-monotonic logic like defeasible logic to reason about intentions in BDI. As such, the research captured in this thesis is theoretical in nature and situated at the crossroads of various disciplines relevant to Artificial Intelligence (AI). More specifically this thesis makes the following contributions: 1. Combining BDI Logics through fibring/dovetailing: BDI systems modeling rational agents have a combined system of logics of belief, time and intention which in turn are basically combinations of well understood modal logics. The idea behind combining logics is to develop general techniques that allow to produce combinations of existing and well understood logics. To this end we adopt Gabbay's fibring/dovetailing technique to provide a general framework for the combinations of BDI logics. We show that the existing BDI framework is a dovetailed system. Further we give conditions on the fibring function to accommodate interaction axioms of the type G [superscript k,l,m,n] ([diamond][superscript k] [superscript l] [phi] [implies] [superscript m] [diamond][superscript n] [phi]) based on Catach's multimodal semantics. This is a major result when compared with other combining techniques like fusion which fails to accommodate axioms of the above type. 2. Extending the BDI framework to accommodate Composite Actions: Taking motivation from a recent work on BDI theory, we incorporate the notion of composite actions, [pi]-1; [pi]-2 (interpreted as [pi]-1 followed by [pi]-2), to the existing BDI framework. To this end we introduce two new constructs Result and Opportunity which helps in reasoning about the actual execution of such actions. We give a set of axioms that can accommodate the new constructs and analyse the set of commitment axioms as given in the original work in the background of the new framework. 3. Intention reasoning as Defeasible reasoning: We argue for a non-monotonic logic of intention in BDI as opposed to the usual normal modal logic one. Our argument is based on Bratman's policy-based intention. We show that policy-based intention has a defeasible/non-monotonic nature and hence the traditional normal modal logic approach to reason about such intentions fails. We give a formalisation of policy-based intention in the background of defeasible logic. The problem of logical omniscience which usually accompanies normal modal logics is avoided to a great extend through such an approach.

Prozesse, seien es Geschäfts- oder Produktionsprozesse, sind ständigen Änderungen unterworfen. Für Unternehmen gilt es, sich im Rahmen von Geschäftsprozessen immer wieder neuen Marktgegebenheiten, Gesetzen oder Kunden anzupassen. Auch Produktionsprozesse müssen bspw. für die Verarbeitung neuer Materialien zugeschnitten werden. Die vorliegende Arbeit beschreibt deshalb, einen umfassenden Ansatz für den Umgang mit Änderungen bzw. Rekonfigurationen von Workflows zu entwickeln. Dieser zeichnet sich durch zwei Schwerpunkte aus: (1) Vollständige Rekonfiguration aller Workflowperspektiven und (2) eine reflexive autonome Steuerung der Rekonfigurationen.

Pour répondre aux besoins des populations humaines, l'agriculture est de plus en plus intensive, utilisant un très grand nombre d'engrais azotés pour augmenter les rendements. Ces engrais sont utilisés parce que l'azote est l'un des facteurs les plus important et limitant pour la croissance des plantes. L’azote sous forme de nitrate est soumis à des problématiques de pollutions pouvant affecter l’environnement ainsi que la santé humaine. Les défis de l'agriculture de demain sont donc de faire face à une population toujours plus nombreuse, tout en limitant l'impact sur notre environnement. C’est pour cela que la recherche et l’agriculture se questionnent de plus en plus sur l’utilisation d’autres produits comme les biostimulants ou des inhibiteurs, afin de limiter les intrants tout en conservant un taux de productivité viable. Une solution consisterait à agir sur les microorganismes du sol liés au cycle de l’azote afin de limiter les pertes des agrosystèmes en azote via le dégagement de gaz à effet de serre (N2O), de lessivage ou/et de volatilisation. En effet, dans les sols, le nitrate est également utilisé par les bactéries dites dénitrifiantes qui le réduisent en N2O (gaz à effet de serre) et N2, représentant alors une perte d’azote pour les cultures et une augmentation de la pollution atmosphérique. Par conséquent, les plantes sont en compétition directe avec ces bactéries pour l'assimilation du nitrate. Récemment, l’équipe encadrante de cette thèse a mis en évidence une stratégie développée par certaines plantes consistant en la production de métabolites secondaires : les procyanidines, qui inhibent la dénitrification des communautés microbiennes du sol. Les procyanidines ont la capacité d’inhiber, chez les bactéries dénitrifiantes, la première étape de la dénitrification transformant le nitrate en nitrite et ainsi d’empêcher l’utilisation du nitrate, sans toutefois exercer un effet antibactérien. Cette stratégie permet de conserver par conséquent le nitrate dans le sol, celui-ci pouvant alors être utilisé par les plantes pour leur nutrition et leur croissance. L’objectif principal de cette thèse a été d’évaluer l’effet de l’application de procyanidines exogènes à plusieurs concentrations en champs sur un modèle de plante d’intérêt économique, la laitue, ainsi que sur différents types de sols. Ces effets ont également été testés sur un autre modèle d’intérêt économique consommatrice d’azote, le céleri. Au cours de ces expérimentations, des mesures ont été effectuées sur (i) l’activité microbienne de dénitrification, (ii) les traits végétaux en lien avec la croissance et (iii) l’abondance des communautés bactériennes dénitrifiantes. Nos résultats montrent une induction d'un BDI en champs, une conservation du nitrate induisant à son tour une amélioration de la croissance des végétaux et une contre sélection par la plante des dénitrifiants. L’autre point abordé, plutôt fondamental mais qui a moins abouti faute de temps, consistait à mettre en évidence via l’utilisation de mutants d’Arabidopsis thaliana affectés dans la production des procyanidines ou surproduisant les procyanidines, un retour vers la plante de l’azote détourné suite au BDI / To meet the needs of human populations, agriculture is increasingly intensive, using a very large number of nitrogen fertilizers to increase yields. These fertilizers are used because nitrogen is one of the most important and limiting factors for plant growth. Nitrogen in the form of nitrate is subject to pollution problems that can affect the environment and human health. The challenges for tomorrow's agriculture are therefore to face an ever-increasing population, while limiting the impact on our environment. This is why research and agriculture are increasingly questioning the use of other products such as biostimulants or inhibitors, in order to limit inputs while maintaining a viable productivity rate. One solution would be to act on soil microorganisms linked to the nitrogen cycle in order to limit nitrogen losses from agrosystems through greenhouse gas (N2O) emissions, leaching and/or volatilization. Indeed, in soils, nitrate is also used by so-called denitrifying bacteria, which reduce it to N2O (greenhouse gases) and N2, representing a loss of nitrogen for crops and an increase in air pollution. As a result, plants compete directly with these bacteria for the assimilation of nitrate. Recently, the supervising team of this thesis has highlighted a strategy developed by some plants consisting of the production of secondary metabolites: procyanidins, which inhibit the denitrification of soil microbial communities. Procyanidins have the ability to inhibit the first step of denitrification in denitrifying bacteria, transforming nitrate into nitrite and thus preventing the use of nitrate, without however exerting an antibacterial effect. This strategy therefore preserves the nitrate in the soil, which can then be used by plants for their nutrition and growth. The main objective of this thesis was to evaluate the effect of the application of exogenous procyanidins at several field concentrations on a plant model of economic interest, lettuce, as well as on different soil types. These effects have also been tested on another model of economic interest that consumes nitrogen, celery. During these experiments, measurements were made on (i) microbial denitrification activity, (ii) plant traits related to growth and (iii) the abundance of denitrifying bacterial communities. Our results show an induction of a BDI in the field, a conservation of nitrate inducing in turn an improvement in plant growth and a counter-selection by the plant of denitrifiers. The other point addressed, which was rather fundamental but less successful due to a lack of time, was to highlight, via the use of Arabidopsis thaliana mutants affected in the production of procyanidins or overproducing procyanidins, a return to the plant of nitrogen diverted following the BDI