Teses / dissertações sobre o tema "Comportement des agents"
Crie uma referência precisa em APA, MLA, Chicago, Harvard, e outros estilos
Veja os 50 melhores trabalhos (teses / dissertações) para estudos sobre o assunto "Comportement des agents".
Ao lado de cada fonte na lista de referências, há um botão "Adicionar à bibliografia". Clique e geraremos automaticamente a citação bibliográfica do trabalho escolhido no estilo de citação de que você precisa: APA, MLA, Harvard, Chicago, Vancouver, etc.
Você também pode baixar o texto completo da publicação científica em formato .pdf e ler o resumo do trabalho online se estiver presente nos metadados.
Veja as teses / dissertações das mais diversas áreas científicas e compile uma bibliografia correta.
Mihoub, Alaeddine. "Apprentissage statistique de modèles de comportement multimodal pour les agents conversationnels interactifs". Thesis, Université Grenoble Alpes (ComUE), 2015. http://www.theses.fr/2015GREAT079/document.
Texto completo da fonteFace to face interaction is one of the most fundamental forms of human communication. It is a complex multimodal and coupled dynamic system involving not only speech but of numerous segments of the body among which gaze, the orientation of the head, the chest and the body, the facial and brachiomanual movements, etc. The understanding and the modeling of this type of communication is a crucial stage for designing interactive agents capable of committing (hiring) credible conversations with human partners. Concretely, a model of multimodal behavior for interactive social agents faces with the complex task of generating gestural scores given an analysis of the scene and an incremental estimation of the joint objectives aimed during the conversation. The objective of this thesis is to develop models of multimodal behavior that allow artificial agents to engage into a relevant co-verbal communication with a human partner. While the immense majority of the works in the field of human-agent interaction (HAI) is scripted using ruled-based models, our approach relies on the training of statistical models from tracks collected during exemplary interactions, demonstrated by human trainers. In this context, we introduce "sensorimotor" models of behavior, which perform at the same time the recognition of joint cognitive states and the generation of the social signals in an incremental way. In particular, the proposed models of behavior have to estimate the current unit of interaction ( IU) in which the interlocutors are jointly committed and to predict the co-verbal behavior of its human trainer given the behavior of the interlocutor(s). The proposed models are all graphical models, i.e. Hidden Markov Models (HMM) and Dynamic Bayesian Networks (DBN). The models were trained and evaluated - in particular compared with classic classifiers - using datasets collected during two different interactions. Both interactions were carefully designed so as to collect, in a minimum amount of time, a sufficient number of exemplars of mutual attention and multimodal deixis of objects and places. Our contributions are completed by original methods for the interpretation and comparative evaluation of the properties of the proposed models. By comparing the output of the models with the original scores, we show that the HMM, thanks to its properties of sequential modeling, outperforms the simple classifiers in term of performances. The semi-Markovian models (HSMM) further improves the estimation of sensorimotor states thanks to duration modeling. Finally, thanks to a rich structure of dependency between variables learnt from the data, the DBN has the most convincing performances and demonstrates both the best performance and the most faithful multimodal coordination to the original multimodal events
Akakzia, Ahmed. "Teaching Predicate-based Autotelic Agents". Electronic Thesis or Diss., Sorbonne université, 2022. https://accesdistant.sorbonne-universite.fr/login?url=https://theses-intra.sorbonne-universite.fr/2022SORUS415.pdf.
Texto completo da fonteAs part of the quest for designing embodied machines that autonomously explore their environments, discover new behaviors and acquire open-ended repertoire of skills, artificial intelligence has been taking long looks at the inspiring fields of developmental psychology and cognitive sciences which investigate the remarkable continuous and unbounded learning of humans. This gave birth to the field of developmental robotics which aims at designing autonomous artificial agents capable of self-organizing their own learning trajectories based on their intrinsic motivations. It bakes the developmental framework of intrinsically motivated goal exploration processes (IMGEPs) into reinforcement learning (RL). This combination has been recently introduced as autotelic reinforcement learning, where autotelic agents are intrinsically motivated to self-represent, self-organize and autonomously learn about their own goals. Naturally, such agents need to be endowed with good exploration capabilities as they need to first physically encounter a certain goal in order to take ownership of and learn about it. Unfortunately, discovering interesting behavior is usually tricky, especially in hard exploration setups where the rewarding signals are parsimonious, deceptive or adversarial. In such scenarios, the agents’ physical situatedness-in the Piagetian sense of the term-seems insufficient. Luckily, research in developmental psychology and education sciences have been praising the remarkable role of socio-cultural signals in the development of human children. This social situatedness-in the Vygotskyan sense of the term-enhances the toddlers’ exploration capabilities, creativity and development. However, deep \rl considers social interactions as dictating instructions to the agents, depriving them from their autonomy. This research introduces \textit{teachable autotelic agents}, a novel family of autonomous machines that can learn both alone and from external social signals. We formalize such a family as a hybrid goal exploration process (HGEPs), where autotelic agents are endowed with an internalization mechanism to rehearse social signals and with a goal source selector to actively query for social guidance. The present manuscript is organized in two parts. In the first part, we focus on the design of teachable autotelic agents and attempt to leverage the most important properties that would later serve the social interaction. Namely, we introduce predicate-based autotelic agents, a novel family of autotelic agents that represent their goals using spatial binary predicates. These insights were based on the Mandlerian view on the prelinguistic concept acquisition suggesting that toddlers are endowed with some innate mechanisms enabling them to translate spatio-temporal information into an iconic static form. We show that the underlying semantic representation plays a pivotal role between raw sensory inputs and language inputs, enabling the decoupling of sensorimotor learning and language grounding. We also investigate the design of such agents' policies and state-action value functions, and argue that combining Graph Neural Networks (GNNs) with relational predicates provides a light computational scheme to transfer efficiently between skills. In the second part, we formalize social interactions as a goal exploration process. We introduce Help Me Explore (HME), a novel social interaction protocol where an expert social partner progressively guides the learning agent beyond its zone of proximal development (ZPD). The agent actively selects to query its social partner whenever it estimates that it is not progressing enough alone. It eventually internalizes the social signals, becomes less dependent on its social partner and maximizes its control over its goal space
Ganna, Mohamed. "Gestion autonome du provisionnement de services en environnnements pervasifs". Paris 6, 2005. http://www.theses.fr/2005PA066504.
Texto completo da fonteKonté, Mamadou. "Investisseurs et Marchés Financiers : du comportement des agents à la formation de prix d'équilibre". Phd thesis, Université Panthéon-Sorbonne - Paris I, 2010. http://tel.archives-ouvertes.fr/tel-00618863.
Texto completo da fonteLe, Cornu François. "Du comportement de stockage des agents économiques à l'analyse macroéconomique des variations de stocks". Montpellier 1, 1991. http://www.theses.fr/1991MON10021.
Texto completo da fonteEconomic activity is a field in which many different types of behavior may be observed concerning inventories. Firms have a multiplicity of reasons for holding inventories (transaction, precaution, speculation, production smoothing, buffer stocks). These reasons lead firms to try to maintain a stable relationship between inventories and sales. However, this classical view is partially questioned by production pull systems (production without inventories). As for households, in addition to classical reasons, they gave more spychological motives for holding inventories (accumulation, safety). Through the specification of macroeconomic models (accelerator, production smoothing), the study of individual inventory behavior leads us to determine the impact of changes in inventories on economic equilibrium
ROLDAN, ASSAD RAMIZ. "Comportement des tensioactifs ioniques en electrophorese capillaire : utilisation comme agents de surface et separations analytiques". Paris 6, 1994. http://www.theses.fr/1994PA066250.
Texto completo da fonteTran, Trung Hau. "Approches évolutionnaires pour le comportement adaptatif d'entités autonomes". Toulouse 3, 2007. http://thesesups.ups-tlse.fr/92/.
Texto completo da fonteThe behavioral simulation of virtual entities in dynamic and unknown environments becomes a complex problem. Deliberative agents cannot handle unforeseen situations. Reactive agents interact with the environment in real time, but they do not easily adapt their behaviors to dynamic environments. The research of this thesis focuses on mechanisms inspired from Artificial Life, offering characteristics of reactivity, adaptation and evolution. We propose an architecture mixing different mechanisms in order to increase the reasoning capacity of virtual entities. This behavioral architecture consists of two models. The first one allows the autonomous entities to act quickly on the environment. This model is based on the potential fields' method which has already proved its efficiency in navigation problems for autonomous entities. However, this solution is expensive because the designer has to define manually the potential fields' parameters for each environment. To solve this problem, we have developed an additional model allowing the autonomous entities to obtain adaptive behaviors in dynamic environments. The second model is an extension of the classifier system XCS, whose role is to evolve discrete actions corresponding to the potential fields' parameters. We have validated our behavioral architecture in a predator-prey problem. The performance of our architecture entirely depends on classifier systems. The possible actions remain discrete and predefined. .
Pauzin, Marie-Christine. "Modélisation du comportement dynamique d'un agent de contraste ultrasonore". Phd thesis, Université de Provence - Aix-Marseille I, 2009. http://tel.archives-ouvertes.fr/tel-00430076.
Texto completo da fontePimenta, Paulo. "Application of model-driven engineering to multi-agent systems : a language to model behaviors of reactive agents". Thesis, Montpellier, 2017. http://www.theses.fr/2017MONTS031/document.
Texto completo da fonteMany users of multi-agent systems (MAS) are very commonly discouraged to model and simulate using current MAS platforms. More specifically, modeling the dynamics of a system (in particular the agent's behaviors) is very often a challenge to users of MAS. That issue is more often observed in the domain of socio-ecological systems (SES), because SES domain experts are rarely programmers. Indeed,the majority of MAS platforms were not conceived taking into consideration domain-experts that are non-programmers. Most of the current MAS tools are not dedicated to SES, or they do not possess an easily understandable formalism to represent behaviors of agents. Moreover, because it is platform-dependent, a model realized in a MAS platform cannot be properly used in another platform due to incompatibility between MAS platforms. To overcome these limitations, we propose a domain-specific language (DSL) to describe the behaviors of reactive agents, regardless of the MAS platform used for simulation. To achieve that, we applied model-driven engineering (MDE), an approach that provides tools to develop DSLs from a meta-model (abstract syntax), textual editors with syntax highlighting (for the concrete syntax) and code generation capabilities (for source-code generation of a model). As a result, we implemented a language and a textual editor that allows SES domain experts to describe behaviors in three different ways that are closed to their natural expression: as equations when they are familiar to those, as a sequence of activities close to natural language or as an activity diagram to represent decisions and a sequence of behaviors using a graphic formalism. To show the generality we also developed code generators targeting two different MAS platforms (Cormas and Netlogo). We tested the code generators by implementing two SES models with the developed DSL. The generated code was targeted for both MAS platforms (Cormas and Netlogo), and successfully simulated in one of them.We conclude that the MDE approach provides adequate tools to develop DSL and code generators to facilitate MAS modeling and simulation by non-programmers. Concerning the developed DSL, although the behavioral aspect of MAS simulation is part of the complexity of modeling in MAS, there are still other essential aspects of model and simulation of MAS that are yet to be explored, such as model's initialization and points of view on the model's simulated world
Edey, Gamassou Claire. "Stress et implication des agents territoriaux : une approche en terme de ressources". Paris 1, 2006. http://www.theses.fr/2006PA010016.
Texto completo da fonteHartmann, Nadine. "Distribution cérébrale d'agents gabaergiques par des systèmes de libération chimique et conséquences comportementales". Nancy 1, 1994. http://docnum.univ-lorraine.fr/public/SCD_T_1994_0445_HARTMANN.pdf.
Texto completo da fonteReynaud, Quentin. "Une architecture hybride et flexible pour agents virtuels en environnement urbain : problématiques de la composition de comportements et de l'anticipation". Electronic Thesis or Diss., Paris 6, 2014. http://www.theses.fr/2014PA066601.
Texto completo da fonteThis thesis deals about human behaviors simulation in an urban context. We focus on the behavior believability (as judged as external observers). That is why our agent have some anticipatory skills, which allow them to use predictions about their near future. Our architecture is a hybrid one, which is very innovative because of its functioning with « high-level modules », which are seen as black-box from the rest of the architecture. Their number and intern modeling are completely free. This makes our architecture very modular and generic, and it is important because the urban simulation domain has many different applications (urbanism, video games, security, etc.), with different constraints. However, this genericity brings another problem, which is the integration of several heterogeneous behavior into the same decisional process. This issue is addressed thanks to a behavior composition mechanism. To conclude, we ensure the scaling up of our architecture with the creation of several levels of detail in the agents modeling
Vandendris, Michel. "Etude du comportement in vivo et de la sensibilité aux agents cytotoxiques de la tumeur transplantable RC". Doctoral thesis, Universite Libre de Bruxelles, 1987. http://hdl.handle.net/2013/ULB-DIPOT:oai:dipot.ulb.ac.be:2013/213501.
Texto completo da fonteMasmoudi, Souhir. "Contagion des anticipations des investisseurs sur le marché financier : une approche par les réseaux et les modèles multi-agents". Thesis, Aix-Marseille, 2016. http://www.theses.fr/2016AIXM1084.
Texto completo da fonteWithin a behavioral approach and given the complexity of financial markets, the aim of this thesis is to examine the extent to which directed networks that governs the interaction among investors as well as their mimicking behavior influence their anticipations and the price dynamics. We propose an artificial asset market populated by chartists and fundamentalists who are allowed to switch from one trading strategy to the other according to their relative performances. Firstly, we study a fully connected network to test for a global interaction. We find that our benchmark model accounts for the emergence of excess volatility of asset prices when chartists dominate the market. Secondly, we restrict our focus to local interactions between investors. We generate a family of network structures that spans regular network, small world network and random network. Thirdly, we introduce a new model that allows us to control (1) the direction of the rewiring process of the links; (2) the randomness of the network; and (3) the asymmetry in its degree distribution by assuming that there are two classes of agents: stars and non-stars. We show that unlike the degree of the randomness of the network, the asymmetry in the degree distribution produces opposite effects depending on whether the network is outward or inward rewired. Finally, we address the question as to how this analysis can be used to produce realistic market dynamics. We find that the presence of a mimicking threshold with a high reaction coefficient provides a better approximation to the characteristics of the distribution of real returns and reproduces the most important stylized facts observed in financial time series
Gadilhe, Agnès. "Comportement aéraulique des enveloppes de bâtiment : détermination numérique des pressions en façade : Modélisation de la perméabilité à l'air". Lyon, INSA, 1990. http://www.theses.fr/1990ISAL0023.
Texto completo da fonte[Pressure models provide an attractive tool for the evaluation of the building ventilation. However, further development and validation of these models requires a better understanding of two critical issues : the pressure distribution around the building and the hydrodynamic behaviour of the envelope. The objectives of this thesis are to present a feasibility study of the k-e model for the evaluation of the pressure distribution and to better understand the flow characteristics through the envelope. First, we compare the pressure coefficients on a seven story building obtained by three existing methods : wind tunnel analysis, on site measurements, numerical approach. The comparison of the results of the numerical simulation with the wind tunnel test on two buildings demonstrates the limits and the difficulties of this promising numerical technique. Then we focus on the behaviour of the envelope - interface between inside and outside. A bibliography review and an on site experiment allow us to identify the type, the location and the flow equation of the various air paths. In order to improve our understanding of the flow through every specific building component, we test in our laboratory apparatus the behaviour of cracks and slots. The limits of the pressurization technique and its associated logarithmic law are clarified. ]
Paraschiv, Corina. "Gestion des agents intelligents sur l'Internet : L'apport conjoint du marketing et de l'économie". Cachan, Ecole normale supérieure, 2002. http://www.theses.fr/2002DENS0039.
Texto completo da fonteDaviet, Stephane. "Etude du comportement humain grâce à la simulation multi-agents et aux méthodes de fouille de données temporelles". Phd thesis, Université de Nantes, 2009. http://tel.archives-ouvertes.fr/tel-00482642.
Texto completo da fonteDaviet, Stéphane. "Étude du comportement humain grâce à la simulation multi-agents et aux méthodes de fouille de données temporelles". Nantes, 2009. http://www.theses.fr/2009NANT2018.
Texto completo da fonteSimulation tools has already proven its efficiency for studying numerous phenomena, but simulating human behavior is still a challenge for both computer and human sciences. In this context, we have worked on the simulation of individual demeanor and social interactions in human groups to study the emergence of collective beahvioral patterns, mixing techniques of artificial intelligence, multi-agent systems, psychological sciences and spatio-temporal data mining. This thesis presents a new model of emotional agent: the Emotion, Feeling, Temperament agent (EFT). Based on BDI architecture, our model integrates the OCC emotional model and the PerformanSe behavioral model. We also present a concrete implementation of our model: the simulation of brain-damaged people’s behavior on a production line. Then, we describe the modeling of the interactions of this system with AgentUML. We mine data produced by the simulation through knowledge discovery techniques applied to temporal data
Salek, Mohammed. "Statistical analysis and modeling of the opening and closing auctions of financial markets". Electronic Thesis or Diss., université Paris-Saclay, 2024. http://www.theses.fr/2024UPAST045.
Texto completo da fonteThis thesis is devoted to the study of opening and closing auctions in European markets with a specific emphasis on the Paris stock exchange. Auctions serve as an essential mechanism to open and close trading days in an orderly way. In particular, closing prices are of crucial importance for both investors and regulators. In contrast to the abundant literature on the continuous trading phase, work on auctions is scarce, partly due to the difficulty in acquiring high-quality data. This thesis is based on a high-quality dataset allowing us to provide novel insights into auction phases, reconstruct their tick-by-tick dynamics, and propose accurate models of the observed phenomena.First, we examine price impact in auctions. We provide a mathematical framework for auctions, compute the average shape of the limit order book at auction time, and its breakdown by agents' latency and account type. We investigate price impact at auction time and address the effect of random clearing times. Second, we adapt a continuous price-time model to the specifics of equity auctions. We show that time-dependent solutions can be obtained in closed-form formulas. We numerically solve the full equations and fit the average dynamics of the limit order book in closing auctions. We conclude by investigating the causes behind the indicative price sub-diffusivity
Darakdjian, Quentin. "Prédiction des performances énergétiques des bâtiments avec prise en compte du comportement des usagers". Thesis, La Rochelle, 2017. http://www.theses.fr/2017LAROS015/document.
Texto completo da fonteContinuous improvement of the building energy performance is associated with the development of increasingly efficient and accurate numerical tools. While the consideration of phenomena related to buildings, systems and weather is well mastered, occupants’ behaviours are modelled in a very simplified way by repetitive scenarios and deterministic laws. The impact of occupants on energy consumption in high-performance buildings is dominant, as evidenced by the recurring gaps between predicted and measured results. The thesis demonstrates, via a multi-agent platform and stochastic models, an update on the ability to model occupants’ presence, their behaviours on windows, occultation devices, artificial lighting and heating setpoint temperatures. The application of the platform applies to office and residential buildings, for new builds and refurbishments. Occupants’ behaviour models are ideally obtained from in situ surveys, laboratory studies or sociological works. The suggested platform is then co-simulated with the EnergyPlus software, to study the influence of the models on a buildings energy performance. In the perspective of energy performance guarantees, this work contributes to the updating and reliability of prediction tools
Reynaud, Quentin. "Une architecture hybride et flexible pour agents virtuels en environnement urbain : problématiques de la composition de comportements et de l'anticipation". Thesis, Paris 6, 2014. http://www.theses.fr/2014PA066601/document.
Texto completo da fonteThis thesis deals about human behaviors simulation in an urban context. We focus on the behavior believability (as judged as external observers). That is why our agent have some anticipatory skills, which allow them to use predictions about their near future. Our architecture is a hybrid one, which is very innovative because of its functioning with « high-level modules », which are seen as black-box from the rest of the architecture. Their number and intern modeling are completely free. This makes our architecture very modular and generic, and it is important because the urban simulation domain has many different applications (urbanism, video games, security, etc.), with different constraints. However, this genericity brings another problem, which is the integration of several heterogeneous behavior into the same decisional process. This issue is addressed thanks to a behavior composition mechanism. To conclude, we ensure the scaling up of our architecture with the creation of several levels of detail in the agents modeling
Militaru, Dorin. "Technologies Internet, systèmes de recommandations et agents intelligents". Paris, ENSAM, 2004. http://www.theses.fr/2004ENAM0036.
Texto completo da fonteThe spread of information technologies (IT) and the growth of the electronic commerce modified the way in which the companies function, forcing them to adopt flexible structures and to produce more efficiently. This research is interested in the part played by new technologies in the setting-up of recommendation systems and information search systems which tend on many markets to control the commercial trades. More precisely, our main objective is to contribute to a better understanding of the recommendation systems by comparing physical economy and electronic commerce. This thesis can contribute to the comprehension of the changes generated in this field by the use of the electronic commerce and to the development of “intelligent” substitutes to the processes currently used. By doing this, we contribute to the emergence of a new face of electronic commerce on Internet by identifying a series of psychological and economic variables which play an important part in the manner in which the economic agents, in particular the companies, act and react. Our main objectives throughout this thesis are to answer to the following questions: Which is the part played by the recommendation systems in the formation of preferences and which is the efficiency of this type of systems? Do the specific characteristics of Internet as an economic environment modify the answers to the previous question? Are the competitiveness factors of the companies on the “new economy” different from those of the traditional economy? Which are the opportunities associated with the electronic commerce, in particular through the “shopbots” which tend to become a “hard” part of the recommender systems?
Klein, François. "Contrôle d'un Système Multi-Agents Réactif par Modélisation et Apprentissage de sa Dynamique Globale". Phd thesis, Université Nancy II, 2009. http://tel.archives-ouvertes.fr/tel-00432354.
Texto completo da fonteLecerf, Ugo. "Robust learning for autonomous agents in stochastic environments". Electronic Thesis or Diss., Sorbonne université, 2022. http://www.theses.fr/2022SORUS253.
Texto completo da fonteIn this work we explore data-driven deep reinforcement learning (RL) approaches for an autonomous agent to be robust to a navigation task, and act correctly in the face of risk and uncertainty. We investigate the effects that sudden changes to environment conditions have on autonomous agents and explore methods which allow an agent to have a high degree of generalization to unforeseen, sudden modifications to its environment it was not explicitly trained to handle. Inspired by the human dopamine circuit, the performance of an RL agent is measured and optimized in terms of rewards and penalties it receives for desirable or undesirable behaviour. Our initial approach is to learn to estimate the distribution of expected rewards from the agent, and use information about modes in this distribution to gain nuanced information about how an agent can act in a high-risk situation. Later, we show that we are able to achieve the same robustness objective with respect to uncertainties in the environment by attempting to learn the most effective contingency policies in a `divide and conquer' approach, where the computational complexity of the learning task is shared between multiple policy models. We then combine this approach with a hierarchical planning module which is used to effectively schedule the different policy models in such a way that the collection of contingency plans is able to be highly robust to unanticipated environment changes. This combination of learning and planning enables us to make the most of the adaptability of deep learning models, as well as the stricter and more explicit constraints that can be implemented and measured by means of a hierarchical planner
Legenvre, Hervé. "Les facultés de l'inventeur : une analyse économique du comportement des inventeurs dans l'incertitude". Paris 11, 2009. http://www.theses.fr/2009PA111003.
Texto completo da fonteMasmoudi, Souhir. "Contagion des anticipations des investisseurs sur le marché financier : une approche par les réseaux et les modèles multi-agents". Electronic Thesis or Diss., Aix-Marseille, 2016. http://www.theses.fr/2016AIXM1084.
Texto completo da fonteWithin a behavioral approach and given the complexity of financial markets, the aim of this thesis is to examine the extent to which directed networks that governs the interaction among investors as well as their mimicking behavior influence their anticipations and the price dynamics. We propose an artificial asset market populated by chartists and fundamentalists who are allowed to switch from one trading strategy to the other according to their relative performances. Firstly, we study a fully connected network to test for a global interaction. We find that our benchmark model accounts for the emergence of excess volatility of asset prices when chartists dominate the market. Secondly, we restrict our focus to local interactions between investors. We generate a family of network structures that spans regular network, small world network and random network. Thirdly, we introduce a new model that allows us to control (1) the direction of the rewiring process of the links; (2) the randomness of the network; and (3) the asymmetry in its degree distribution by assuming that there are two classes of agents: stars and non-stars. We show that unlike the degree of the randomness of the network, the asymmetry in the degree distribution produces opposite effects depending on whether the network is outward or inward rewired. Finally, we address the question as to how this analysis can be used to produce realistic market dynamics. We find that the presence of a mimicking threshold with a high reaction coefficient provides a better approximation to the characteristics of the distribution of real returns and reproduces the most important stylized facts observed in financial time series
Elleuch, Fatma. "Transférabilité d'une modélisation-simulation multi-agents : le comportement inter-gares des voyageurs de la SNCF lors des échanges quai-train". Thesis, Paris, CNAM, 2019. http://www.theses.fr/2019CNAM1227/document.
Texto completo da fonteThis thesis focuses on the definition and the validation of a crowd-simulation modeling method. This method allows transferring principles and measurements established in a source observation context to a target one. From the observation of pedestrian behavior in a social and urban context, our approach infers rules and motivations in order to model and simulate the individual pedestrian behavior in this source context. This allows, in a second step, to evaluate the transferability of the model to a given target context. Our work is based on a real case: the simulation of the passenger’s behavior while boarding and alighting trains (platform-train exchange) in a high-traffic station (dense situation). Starting from the state of the art of theories, models and types of simulation, we choose an approach based on the microscopic simulation of social agent’s motivation. This choice relies mainly on the fact that, compared to a classic crowd simulation, it has, a priori, a better transferability potential. The goal of this type of simulation is to study the architectural evolutions of the stations that could allow to control or shorten train dwell times
Elleuch, Fatma. "Transférabilité d'une modélisation-simulation multi-agents : le comportement inter-gares des voyageurs de la SNCF lors des échanges quai-train". Electronic Thesis or Diss., Paris, CNAM, 2019. http://www.theses.fr/2019CNAM1227.
Texto completo da fonteThis thesis focuses on the definition and the validation of a crowd-simulation modeling method. This method allows transferring principles and measurements established in a source observation context to a target one. From the observation of pedestrian behavior in a social and urban context, our approach infers rules and motivations in order to model and simulate the individual pedestrian behavior in this source context. This allows, in a second step, to evaluate the transferability of the model to a given target context. Our work is based on a real case: the simulation of the passenger’s behavior while boarding and alighting trains (platform-train exchange) in a high-traffic station (dense situation). Starting from the state of the art of theories, models and types of simulation, we choose an approach based on the microscopic simulation of social agent’s motivation. This choice relies mainly on the fact that, compared to a classic crowd simulation, it has, a priori, a better transferability potential. The goal of this type of simulation is to study the architectural evolutions of the stations that could allow to control or shorten train dwell times
Dia, Nada. "Nanoparticules photomagnétiques et agents de contraste à base de réseaux de coordination". Phd thesis, Université Paris Sud - Paris XI, 2013. http://tel.archives-ouvertes.fr/tel-00813757.
Texto completo da fonteRey, Jean-Pierre. "Etude de quelques effets du comportement du manager sur les conduites de jeunes joueurs et joueuses de hand-ball : une approche quasi-expérimentale". Clermont-Ferrand 2, 1990. http://www.theses.fr/1990CLF21508.
Texto completo da fonteHanon, David. "Modèle décisionnel orienté comportement fondé sur le vote : Application à la navigation d'agents autonomes en environnement simulé". Phd thesis, Université de Valenciennes et du Hainaut-Cambresis, 2006. http://tel.archives-ouvertes.fr/tel-00135931.
Texto completo da fonteDarty, Kevin. "Évaluation de la qualité des comportements des agents en simulation : application à un simulateur de conduite en environnement virtuel". Thesis, Paris 6, 2015. http://www.theses.fr/2015PA066339/document.
Texto completo da fonteThis thesis is in the context of the Multi-Agents Simulation and is interested in evaluating the ability of agents to reproduce human behaviors. This problem appears in many domains such as Virtual Reality and Embodied Conversational Agents. The dominant approach to evaluate these behaviors uses Social Sciences questionnaires. There are only few approaches based on Artificial Intelligence and automatic data analysis at the microscopic scale. We show in this thesis that the evaluation of behavior can benefit from both approaches when used jointly. First, we present a method for evaluating the agents' behavior quality. It combines the Artificial Intelligence approach and the Social Science approach. The first one is based on simulation logs clustering. The second one evaluates the users by an annotation of the behaviors. We then present an algorithm that compare agents to humans in order to assess the capacities, the lacks, and the errors in the agent model, and provide metrics. We then make these behaviors explicite based on user categories. Finally, we present a cycle for automatic calibration of the agents and an exploration of the parameter space. Our evaluation method is usable for the analysis of an agent model, and for comparing several agent models. We applied this methodology on several driver behavior studies to analyse the road traffic simulation ARCHISIM, and we present the obtained results
Aloui, Madiha. "Élaboration et fabrication de films nanocomposites à matrice biosourcée pour application textile à usage unique : étude du comportement à la rupture". Thesis, Lille 1, 2011. http://www.theses.fr/2011LIL10190.
Texto completo da fonteThe aim of this work is the development and the behavior analysis of blown nanocomposite films based on polylactic acid (PLA), obtained by extrusion-blowing for single-use textile application. The use of the PLA is justified by its natural origin and by its good physical properties. However, PLA presents some limits (in particular, in terms of sensibility to the thermal degradation, processabilty by extrusion-blowing and flexibility) requiring its modification by branching and plasticization. Te efficiency of these modifications was estimated first of all. Adding Nanoclay is justified by the barrier properties aimed. Two nanoclays, associated to different surface treatment, were incorporated to PLA by melt intercalation using an extruder. The dispersion level of the modified clays was analyzed by global techniques (rheology, scanning electron microscopy) or local techniques (X-ray diffraction, transmission electron microscopy) and leads to characterize the efficiency of clay treatments. Nanocomposites so elaborated were used for manufacturing films by extrusion-blowing. The characterization of their mechanical and barrier properties allowed to optimize the film composition considering the aimed application. Finally, film tenacity was studied by the essential work of fracture method for blown films. The relevance of this method is studied via the validation of hypotheses necessary for its application. The technique that measures the displacement field has allowed us to study two of these hypotheses
Ha, Anh Tung. "Effets des agents interfaciaux copolymères tribloc sur le comportement en rupture et en écoulement du mélange de polystyrène/caoutchouc-éthylène-propylène". Mémoire, Université de Sherbrooke, 1999. http://savoirs.usherbrooke.ca/handle/11143/1094.
Texto completo da fonteHa, Anh Tung. "Effets des agents interfaciaux copolymères tribloc sur le comportement en rupture et en écoulement du mélange de polystyrène/caoutchouc-éthylène-propylène". Sherbrooke : Université de Sherbrooke, 2000.
Encontre o texto completo da fonteBangalore, Kantharaju Reshmashree. "Modelling Cohesive Behaviours for Virtual Agents in Multiparty Interactions". Electronic Thesis or Diss., Sorbonne université, 2021. http://www.theses.fr/2021SORUS230.
Texto completo da fonteGroup interactions are a commonly used form of communication among humans. Often the members of a group are involved in discussing, making decisions and exchanging ideas, under different settings (e.g., meeting, conference, party etc.). Group Cohesion describes the shared bond that drives the members to stay together and to want to work together to achieve group goals. In group interactions, humans communicate and coordinate with each other via a number of verbal and nonverbal behaviours. In this research work, as a first step we recognise the relation between group cohesion and certain non-verbal social signals of interest. Next, we present the results on automatic estimation of cohesion levels in groups using different features and feature representation techniques for groups. Virtual agents, a computer-generated animated character with human-like non-verbal behaviours, have been widely used for human-computer interactions in various applications e.g., educational agents, health coaches, training assistants etc. Most of the applications so far have focused on developing agents for dyadic interactions i.e., a single agent and user. A group of agents (multiparty) can be potentially effective in persuading, motivating and educating the users through interactive discussions. In the next step, we develop a multiparty model involving multiple autonomous agents that are capable of displaying cohesive group behaviour i.e., shared commitment to group tasks and positive relationship among the agents. Considering the surge in the range of applications using virtual agents, it is important to study the interactions between multiple agents and the user and understand the effects of using such a system. We hypothesise that the use of a multi-agent system would allow the user to be more engaged in the discussion and provide different perspectives on the same issue and facilitate the users to make informed decisions. Therefore, in the final step we conduct multiple user evaluation studies to understand the effects of multiparty interactions on the user and their perceptions e.g., the level of trust, persuasion. We present insights into the most effective form of interactions for promoting behaviour change or persuading the user using different group conversational topics. To summarise, in this thesis we recognise the association between certain non-verbal social signals and group cohesion, present the estimation accuracy using features extracted from these signals, develop a multiparty model to simulate a cohesive group of agents displaying prominent social signals and finally evaluate the effectiveness of such a model in the context of behaviour change and its effects on user’s perceptions
Darty, Kevin. "Évaluation de la qualité des comportements des agents en simulation : application à un simulateur de conduite en environnement virtuel". Electronic Thesis or Diss., Paris 6, 2015. http://www.theses.fr/2015PA066339.
Texto completo da fonteThis thesis is in the context of the Multi-Agents Simulation and is interested in evaluating the ability of agents to reproduce human behaviors. This problem appears in many domains such as Virtual Reality and Embodied Conversational Agents. The dominant approach to evaluate these behaviors uses Social Sciences questionnaires. There are only few approaches based on Artificial Intelligence and automatic data analysis at the microscopic scale. We show in this thesis that the evaluation of behavior can benefit from both approaches when used jointly. First, we present a method for evaluating the agents' behavior quality. It combines the Artificial Intelligence approach and the Social Science approach. The first one is based on simulation logs clustering. The second one evaluates the users by an annotation of the behaviors. We then present an algorithm that compare agents to humans in order to assess the capacities, the lacks, and the errors in the agent model, and provide metrics. We then make these behaviors explicite based on user categories. Finally, we present a cycle for automatic calibration of the agents and an exploration of the parameter space. Our evaluation method is usable for the analysis of an agent model, and for comparing several agent models. We applied this methodology on several driver behavior studies to analyse the road traffic simulation ARCHISIM, and we present the obtained results
Meskaoui, Nada. "Un environment de modelisation et de simulation des systèmes multi-agent en reseaux de telecommunication : nouvel environnement, outil et comportement". Paris 6, 2004. http://www.theses.fr/2004PA066228.
Texto completo da fonteBiancardi, Béatrice. "Les premières secondes comptent : gérer les premières impressions pour un agent virtuel plus engageant". Electronic Thesis or Diss., Sorbonne université, 2019. https://accesdistant.sorbonne-universite.fr/login?url=https://theses-intra.sorbonne-universite.fr/2019SORUS037.pdf.
Texto completo da fonteThe first moments of an interaction with a virtual character are critical since users form impressions about them, which can affect the rest of the interaction, in terms of engagement and willingness to continue it. In this Thesis we present a computational model for managing user’s impression of agent’s warmth and competence (W&C), the 2 fundamentals dimensions of social cognition. The goal of the agent is to adapt its non-verbal behaviour in real-time during an interaction, according to user’s non-verbal reactions. Our approach starts from the analysis of a corpus of human-human interactions in order to identify a set of non-verbal behaviours eliciting different degrees of W&C. A perceptual study then investigated how these behaviours are perceived when performed by a virtual agent. A reinforcement learning (RL) algorithm has been developed to allow the agent to learn in real-time the behaviours which give the best impression to the user, according to its goal. User’s non-verbal reactions (computed from low-level signals such as facial action units) are used as a reward for the RL. We have personalized the computational model in order to adapt the agent’s behaviours with the goal of maximizing user’s engagement or impressions of agent’s W&C. Two use cases have been run in order to evaluate the impact of an adapting agent on user’s impressions and perception of the interaction, compared to a non-adapting agent. In the first experiment the agent adapted its self-presentational strategies in order to maximise user’s engagement. In the second one the agent learned the best combinations of non-verbal behaviours to display in order to maximise user’s impressions of its W&C
Guyader, Olivier. "Evaluation economique et simulation des effets des regulations sur le comportement des agents : le cas des pecheries de thon germon de l'atlantique nord est". Rennes 1, 1998. http://www.theses.fr/1998REN11015.
Texto completo da fonteKhenfer, Jamel. "Trois essais sur l'influence des agents externes de contrôle sur la poursuite du but du consommateur : de la religion aux marques". Electronic Thesis or Diss., Aix-Marseille, 2015. http://www.theses.fr/2015AIXM1067.
Texto completo da fontePeople are often reminded that influential external entities are by their side while they pursue important personal goals. In the context of reduced personal control, feeling their presence compensate for the realization of one’s own limitations when it comes to producing desired outcomes and avoiding undesired ones. Such reminders make people feel confident that there is structure and order in the world and help them cope with the anxiety inducing fear of randomness and chaos. In other words, when people feel that they do not have control over the pursuit of a goal, perceiving that overall “things are under control” because of the intervention of external agents promote the belief that actions and outcomes are interconnected. Just as religion often reminds believers that a deity is by their side, brands often explicitly remind their customers of their presence for goals they are pursuing. However, one may question the relevance of such a tactic. Prior work has provided extensive evidence showing that promoting influential external agents was detrimental to individuals’ willingness to take goal-directed actions (e.g., reactance and social loafing). In this dissertation, we examine this contradiction. We propose and test a theoretical reconciliation based on three researches. In doing so, we develop the notion of functional homology between religion and brand by arguing and demonstrating that both an interventionist deity and an interventionist brand serve the basic need for order and structure when people deal with a control threat. We further discuss the limitations of this work, directions for further research and the implications for marketing practice
Amblard, Frédéric. "Comprendre le fonctionnement de simulations sociales individus-centrées". Phd thesis, Clermont-Ferrand 2, 2003. http://www.theses.fr/2003CLF21468.
Texto completo da fonteAgbossou, Igor. "Modélisation et simulation multi-agents de la dynamique urbaine : application à la mobilité résidentielle". Phd thesis, Université de Franche-Comté, 2007. http://tel.archives-ouvertes.fr/tel-00924741.
Texto completo da fonteBenson, Marie Anne. "Pouvoir prédictif des données d'enquête sur la confiance". Master's thesis, Université Laval, 2021. http://hdl.handle.net/20.500.11794/69497.
Texto completo da fonteConfidence survey data are time series containting the responses to questions aiming to measure confidence and expectations of economic agents about future economic activity. The richness of these data and their availability in real time attracts the interest of many forecasters who see it as a way to improve their traditional forecasts. In this thesis, I assess the predictive power of survey data for the future evolution of Canadian GDP, while comparing the forecasting performance of the Conference Board of Canada own confidence indices to the indicators I construct using principal component analysis. Using three simple linear models, I carry out an out-of-sample forecasting experiment with rolling windows on the period 1980 to 2019. The results show that principal component analysis provides better-performing indicators than the indices produced by the Conference Board. However, the results of the study cannot show that clear that confidence improves forecasting unambiguently once the lagged growth rate of GDP is added to the analysis.
Khenfer, Jamel. "Trois essais sur l'influence des agents externes de contrôle sur la poursuite du but du consommateur : de la religion aux marques". Thesis, Aix-Marseille, 2015. http://www.theses.fr/2015AIXM1067.
Texto completo da fontePeople are often reminded that influential external entities are by their side while they pursue important personal goals. In the context of reduced personal control, feeling their presence compensate for the realization of one’s own limitations when it comes to producing desired outcomes and avoiding undesired ones. Such reminders make people feel confident that there is structure and order in the world and help them cope with the anxiety inducing fear of randomness and chaos. In other words, when people feel that they do not have control over the pursuit of a goal, perceiving that overall “things are under control” because of the intervention of external agents promote the belief that actions and outcomes are interconnected. Just as religion often reminds believers that a deity is by their side, brands often explicitly remind their customers of their presence for goals they are pursuing. However, one may question the relevance of such a tactic. Prior work has provided extensive evidence showing that promoting influential external agents was detrimental to individuals’ willingness to take goal-directed actions (e.g., reactance and social loafing). In this dissertation, we examine this contradiction. We propose and test a theoretical reconciliation based on three researches. In doing so, we develop the notion of functional homology between religion and brand by arguing and demonstrating that both an interventionist deity and an interventionist brand serve the basic need for order and structure when people deal with a control threat. We further discuss the limitations of this work, directions for further research and the implications for marketing practice
Dermouche, Soumia. "Leveraging the dynamics of non-verbal behaviors : modeling social attitude and engagement in human-agent interaction". Electronic Thesis or Diss., Sorbonne université, 2019. https://accesdistant.sorbonne-universite.fr/login?url=https://theses-intra.sorbonne-universite.fr/2019SORUS271.pdf.
Texto completo da fonteSocial interaction implies exchange between two or more persons, where they adapt their behaviors to each others. With the growing interest in human-agent interactions, it is desirable to make these interactions natural and human like. In this context, we aimed at enhancing the quality of the interaction between users and Embodied Conversational Agents ECAs by (1) endowing the ECA with the capacity to express social attitudes, such as being friendly or dominant depending its role or relationship with its interaction partners; (2) adapting the agent's behavior according to the user's behavior, hence, the conversation partners influence each others through an interaction loop, thus, enhancing the interaction quality; (3) predicting the user's engagement level and adapting the agent's behavior accordingly. We take advantage of the recent advances in machine learning, more specifically, temporal sequence mining and neural networks to model these capacities in the ECA. The first model is used to learn relevant patterns (sequences) of non-verbal signals that best represent attitude variations, and then reproduce them on the agent. The latter is used to encompass the dynamics of non-verbal signals. Two use cases have been explored using the well-known LSTM model: agent's behavior adaptation based on both agent's and user's behavior history, and user's engagement prediction based on his/her own behavior history. The implemented models and algorithms have been validated through a number of perceptive studies as well as through rigorous quantitative analysis of the obtained results. In addition, the realized models have been integrated into a virtual-agent platform
Lacroix, Benoit. "Normer pour mieux varier ? La différenciation comportementale par les normes, et son application au trafic dans les simulateurs de conduite". Phd thesis, Université des Sciences et Technologie de Lille - Lille I, 2009. http://tel.archives-ouvertes.fr/tel-00527846.
Texto completo da fontePerrot, Emilie. "L'activité transférentielle, une ressource au service du développement du métier : le cas des agents d'escale et de service commercial en gare". Thesis, Paris, CNAM, 2017. http://www.theses.fr/2017CNAM1087/document.
Texto completo da fonteBased on an intervention realised in SNCF company for two groups of commercial and station agents, this thesis explores the function of the transferential activity than can be deployed during intervention. We seek to explain the way « the motion of places » which are implemented in the clinical intervention, can be a resource for the development of the job. A multimodale analysis of research materials which mainly come from crossed-confrontation focuses on two types of signs. On the one hand, the presence of a transferential activity indices in the dialogical activity, and the other hand, the signs of the development of the job. This results leads us to present a model of transferential activity. This specificity is her object. Either it assigns the other person (or oneself) a place, or it makes the place change (his or her own or someone else’s one). The transferential activity can be suffered as well as used by the researcher or the professional. It’s the interference of both transferential activities that can be driving force and promote a development of the job on these differents registers. We conclude with the development of the transferential activity and its « functional » nomadism. The transferential activity is not only an activity in its own rightn but it is also the product of another activity. And, finally, it is a psychological instrument to the benefit of the current activity. The perspectives opened by this work concern the transformative potential of the transferential activity. The professional’s activity in the steering committees must be developed
Asadullah, Muhammad Ali. "Comparative perspective of training evaluation practices : a study of ‘Entry Level Professional Training’ of call center agents inside Pakistan". Thesis, Aix-Marseille, 2012. http://www.theses.fr/2012AIXM1064.
Texto completo da fonteHigher turnover, increased recruitment, job complexity and lack of vocational training for call center agents has increased the demand of entry level professional training of call center agents. The question is that how call center professionals determine the value of entry level professional training of call center agents. Call centers can be divided in two groups ‘In-House' and ‘Subcontractor' with respect to the ownership. The objective of this study is to investigate the difference in evaluation of entry level professional training of call center agents among these two groups of call centers. ‘The Kirkpatrick Model' of training evaluation was used to study these differences with addition of 5th level return on investment. Further, we identified stakeholder groups inside call centers. Then we attempted to identify the most important stakeholder groups who may benefit from evaluation information obtained at each level of ‘TKM' model. Data was collected from call centers inside Pakistan by using a mixed methods approach for data collection. Qualitative data was collected through 15 group interviews which were conducted with professionals of 13 call centers inside Pakistan. Quantitative data was collected from almost 203 call center professionals of 90 call centers from three cities Lahore, Karachi and Islamabad. We found that except for first level ‘Reaction' there was no difference in evaluation practices of in-house and subcontractor call centers. Moreover, ‘size' of call center and ‘duration' of entry level professional training had a significant effect on the difference in evaluation practices
Siebert, Julien. "Approche multi-agent pour la multi-modélisation et le couplage de simulations : application à l'étude des influences entre le fonctionnement des réseaux ambiants et le comportement de leurs utilisateurs". Phd thesis, Université Henri Poincaré - Nancy I, 2011. http://tel.archives-ouvertes.fr/tel-00642034.
Texto completo da fonte