Dissertations / Theses on the topic 'Markovien'
Create a spot-on reference in APA, MLA, Chicago, Harvard, and other styles
Consult the top 50 dissertations / theses for your research on the topic 'Markovien.'
Next to every source in the list of references, there is an 'Add to bibliography' button. Press on it, and we will generate automatically the bibliographic reference to the chosen work in the citation style you need: APA, MLA, Harvard, Chicago, Vancouver, etc.
You can also download the full text of the academic publication as pdf and read online its abstract whenever available in the metadata.
Browse dissertations / theses on a wide variety of disciplines and organise your bibliography correctly.
Barbot, Nelly. "Files d'attente fluides en environnement markovien." Rennes 1, 2002. http://www.theses.fr/2002REN10094.
Full textDinetan, Lee. "Ruine et investissement en environnement markovien." Thesis, Toulouse 3, 2015. http://www.theses.fr/2015TOU30141/document.
Full textThis thesis aims at modelling and optimize an agent's (called "he") investment strategies when subjected to a Markovian environment, and to a liquidity risk happening when he runs out of liquid assets during an expense. Throughout this work, we deem that he aims at avoiding default; for this purpose, investment opportunities are available to him, allowing to increase his future expected incomes at the price of an immediate expense, therefore risking premature bankruptcy since investment is deemed illiquid: our goal is to find conditions under which incurring such liquidity risks is more advisable than declining a permanent income
Bouhlel, Nizar. "Caractérisation de texture d'échographie RF par champ markovien." Phd thesis, Université René Descartes - Paris V, 2006. http://tel.archives-ouvertes.fr/tel-00168529.
Full textAMOURA, LAHLOU. "Modele markovien pyramidal flou et segmentation statistique d'images." Paris 6, 1998. http://www.theses.fr/1998PA066011.
Full textBouhlel, Nizar. "Caractérisation de texture d'écographie RF par champ markovien." Paris 5, 2006. http://www.theses.fr/2006PA05S014.
Full textThe ultrasound is a medical imaging tool that imposes itself for the diagnosis of numerous pathologies. As a consequence, many image filed studies are concerned with these images to provide tools of analysis and tissue characterization. The objective of this thesis is to exploit spatial markovien models representing the ultrasound texture that exist in the image to extract susceptible interactions that would describe the organization of these textures. In the first part, we evaluate the efficiency of the distributions proposed in the previous study modeling the RF envelope amplitude. We illustrate, through texture simulations, the links between the parameters of these distributions and the parameters of the scatterers, namely the density, the amplitude and the spacing. In the second part, we elaborate our texture spatial models inspired by the probability modeling RF envelope amplitude. So we obtain in every pixel of the image a local distribution of type K or Nakagami. Simulation and parameter estimation were developed. The third part is dedicated to the application of the spatial models on RF simulated images. We show here the adequate models for the description of the spatial arrangement of the scatterers constituting the tissue, and the connection of the model parameters with the intrinsic properties of the scatterers. Finally, the future development of this approach are to be discussed
Besbes, Ghina. "Modélisation de dialogues à l'aide d'un modèle Markovien caché." Thesis, Université Laval, 2010. http://www.theses.ulaval.ca/2010/26989/26989.pdf.
Full textModeling human-machine dialogue is a research area that encompasses several disciplines such as philosophy, computer science, as well as cognitive and social sciences. It aims to replicate the human ability to learn optimal strategies of dialogue. Furthermore, it aims to design and evaluate management systems for dialogue, and to study the nature of the conversations in more detail. Moreover, few simulation models of existing dialogues were considered good. This thesis presents a hidden Markov model that predicts the action of the user in dialogue systems on the basis of the previous system action. The learning model has been realized through an approach to unsupervised learning using different methods of cross validation. As for model evaluation, it has been done using different metrics. The evaluation results were below expectation. Nonetheless, they are satisfactory compared to previous work. Ultimately, avenues for future research are proposed to overcome this problem. Keywords: natural language processing, spoken dialogue human-machine, Hidden Markov Model (HMM), unsupervised learning, cross validation.
Salzenstein, Fabien. "Modele markovien flou et segmentation statistique non suupervisee d'images." Rennes 1, 1996. http://www.theses.fr/1996REN10194.
Full textProfeta, Christophe. "Pénalisations, pseudo-inverses et peacocks dans un cadre markovien." Thesis, Nancy 1, 2010. http://www.theses.fr/2010NAN10088/document.
Full textAs suggested by the title, this thesis comprises three parts.- The first part is dedicated to the penalization of regular recurrent linear diffusions. More precisely, we start by examining null recurrent diffusions, and we exhibit a large class of functionals for which the penalization principle is satisfied. This study relies on the construction of a sigma-finite measure W similar to that of Najnudel-Roynette-Yor. We then deal with the case of the penalization of a positively recurrent diffusion (reflected on an interval) with an exponential function of its local time at 0. The results we obtain in this set-up are quite different from the null recurrent framework, and we see a new phenomena of composition of penalizations.- In the second part, we extend the notion of pseudo-inverses (a notion recently introduced by Madan-Roynette-Yor in the framework of Bessel processes) to more general diffusions. We show in particular that we may realize the family of pseudo-inverses associated to a diffusion started from 0 and taking positive values as the last passage times of another diffusion, constructed thanks to Biane's transform.- The last part of this thesis deals with peacocks, i.e. with processes which are increasing in the convex order. A theorem due to Kellerer states that to every peacock, one can associate a martingale which has the same one-dimensional marginals. Guided by this theorem, we first exhibit large families of peacocks, essentially constructed from "conditionally monotone" processes, and we then associate martingales to some of these peacocks thanks to the Skorokhod embeddings of Hall-Breiman, Bass and Azéma-Yor
Saint, Pierre Philippe. "Modèles multi-états de type Markovien et application à l'asthme." Phd thesis, Université Montpellier I, 2005. http://tel.archives-ouvertes.fr/tel-00010146.
Full textDevillers, Laurence. "Reconnaissance de parole continue avec un systeme hybride neuronal et markovien." Paris 11, 1992. http://www.theses.fr/1992PA112426.
Full textYe, Yinna. "PROBABILITÉ DE SURVIE D'UN PROCESSUS DE BRANCHEMENT DANS UN ENVIRONNEMENT ALÉATOIRE MARKOVIEN." Phd thesis, Université François Rabelais - Tours, 2011. http://tel.archives-ouvertes.fr/tel-00605751.
Full textBENDJEBBOUR, AZZEDDINE. "Segmentation d'images multisenseur par fusion de dempster-shafer dans un contexte markovien." Paris 6, 2000. http://www.theses.fr/2000PA066037.
Full textHamila, Mohamed amine. "Planification multi-agents dans un cadre markovien : les jeux stochastiques à somme générale." Phd thesis, Université de Valenciennes et du Hainaut-Cambresis, 2012. http://tel.archives-ouvertes.fr/tel-00747848.
Full textIllayk, Abbas. "Évolution du nombre de composants en panne pour un système réparable non-Markovien." Lille 1, 2003. https://ori-nuxeo.univ-lille1.fr/nuxeo/site/esupversions/8a4c0b86-722d-4ea2-8464-13200e65618f.
Full textBoussarsar, Riadh. "Contribution des mesures floues et d'un modèle markovien à la segmentation d'images couleur." Rouen, 1997. http://www.theses.fr/1997ROUES036.
Full textHamila, Mohammed Amine. "Planification multi-agents dans un cadre markovien : les jeux stochastiques à somme générale." Thesis, Valenciennes, 2012. http://www.theses.fr/2012VALE0014/document.
Full textPlanning agent’s actions in a dynamic and uncertain environment has been extensively studied. The framework of Markov decision process provides tools to model and solve such problems. The field of game theory has allowed the study of strategic interactions between multiple agents for a given game. The framework of stochastic games is considered as a generalization of the fields of Markov decision process and game theory. It allows to model systems with multiple agents and multiple states. However, planning in a multi-agent system is considered difficult : agent’s decisions depend not only on its actions but also on actions of the other agents. The work presented in this thesis focuses on decision making in distributed multi-agent systems. Existing works in this field allow the theoretical resolution of stochastic games but place severe restrictions and ignore some crucial problems of the model. We propose a decentralized planning algorithm for the model of stochastic games. Our proposal is based on the Value-Iteration algorithm and on the concept of Nash equilibrium. To improve the resolution process and to deal with large problems, we sought to ease decision making and limit the set of joint actions at each stage. The proposed algorithm was validated on a coordination problem including several agents and various experiments were conducted to assess the quality of the resulting solution
de, Saporta Benoîte. "Etude de la solution stationnaire de l'équation Y(n+1)=a(n)Y(n)+b(n) à coefficients aléatoires." Phd thesis, Université Rennes 1, 2004. http://tel.archives-ouvertes.fr/tel-00007497.
Full textCamescasse, François-Xavier. "Etude femtoseconde de la relaxation des électrons dans les semiconducteurs en régime non-markovien." Phd thesis, Ecole Polytechnique X, 1998. http://pastel.archives-ouvertes.fr/pastel-00005946.
Full textBailey, Ian. "Planification heuristique avec les processus de décision markovien et création d'un environnement de programmation." Mémoire, Université de Sherbrooke, 2005. http://savoirs.usherbrooke.ca/handle/11143/4630.
Full textAltuve, Miguel. "Détection multivariée des épisodes d'apnée-bradycardie chez le prématuré par modèles semi-markovien cachés." Rennes 1, 2011. http://www.theses.fr/2011REN1S053.
Full textThis dissertation studies the early detection of apnea-bradycardia (AB) events in preterm infants. After defining the importance of AB detection from a clinical point of view, a methodological approach is proposed. It relies on a data mining process that includes data cleansing and feature extraction. In chapter 3, a novel method based on evolutionary algorithms, for optimizing the thresholds and the analysis windows, is proposed to adapt the algorithms of the ECG signal to the specific characteristics of preterm infants, very different from the EGC of adult. In chapter 4, a semi-Markovian approach is adapted for modeling of dynamics and several improvements are proposed : heterogeneous models, adaptation to online processing, optimization of experiments, are reported on simulated and read signals. They clearly highlight the importance of considering the dynamic of the signals. They also emphasize that with a suitable pre-treatment such as the quantification of observations and the introduction of delay between the observable, a significant gain in performance can be observed
Camescasse, François-Xavier. "Etude femtoseconde de la relaxation des electrons dans les semiconducteurs en regime non-markovien." Palaiseau, Ecole polytechnique, 1998. http://www.theses.fr/1998EPXX0076.
Full textRekik, Ahmed. "Segmentation statistique et fusion d'images satellitaires par la théorie de l'évidence dans un contexte markovien." Littoral, 2008. http://www.theses.fr/2008DUNK0207.
Full textThe work developed in this thesis, is focused on the unsupervised statistical segmentation of satellite images in a Markovian context, and their fusion through the evidence theory. Indeed we have developed in this work an optimal statistical approach for the segmentation of satellite images, through the integration and the contribution of several algorithms, especially for the initialisation step by using the K-means clustering algorithm for a better definition of the image classes, then we wanted to rectify and standardize these classes through the Markov fields which allowed the consideration of the neighbourhood concept in the classification phase. For the modelling of the different classes of the image, we opted for the Pearson system for its flexibility and its adaptation by offering a range of different and optimal distributions. Finally, concerning the estimation of the different attributes of each class of the image, we used the EM and SEM algorithms. In order to optimize this work, we integrated in our approach an image fusion phase based on the evidence theory (belief function), which allowed a better decision in the segmentation stage, through the exploitation of the number of information present in the multispectral and multi-temporal images
Guerrero, Guillaume. "Implications des changements de régime markovien dans des modèles à anticipations rationnelles : une exploration empirique." Paris 1, 2004. http://www.theses.fr/2004PA010038.
Full textCaillault, Emilie. "Architecture et Apprentissage d'un Système Hybride Neuro-Markovien pour la Reconnaissance de l'Écriture Manuscrite En-Ligne." Phd thesis, Université de Nantes, 2005. http://tel.archives-ouvertes.fr/tel-00084061.
Full textPoisson, Émilie. "Architecture et apprentissage d'un système hybride neuro-markovien pour la reconnaissance de l'écriture manuscrite en-ligne." Nantes, 2005. http://www.theses.fr/2005NANT2082.
Full textThis thesis deals with the study, the conception, the development and the test of an online unconstrained handwriting word recognition system for an omni-writer application. The proposed system is based on a hybrid architecture including on the one hand, a neural convolutional network (TDNN and/or SDNN), and on the other hand Hidden Markov Models (HMM). The neural network has a global vision and works at the character level, while the HMM works on a more local description and allows the extension from the character level to the word level. The system was first dedicated for processing isolated characters (digits, lowercase letters, uppercase letters). This architecture has been optimized in terms of performances and size. The second part of this work concerns the extension to the word level. In this case, we have defined a global training scheme directly at the word level. It allows to insure the global convergence of the system. It relies on an objective function that combines two main criteria: one based on generative models (typically by maximum likelihood estimation) and the second one based on discriminant criteria (maximum mutual information). Several results are presented on MNIST, IRONOFF and UNIPEN databases. They show the influence of the main parameters of the system, either in terms of topologies, information sources, and training models (number of states, criteria weighting, duration)
Caplier, Alice. "Modele markovien de detection de mouvement dans les sequences d'images : approche spatio-temporelle et mises en uvre temps reel." Grenoble INPG, 1995. http://www.theses.fr/1995INPG0170.
Full textNguyen-Van, Quoc-Anh Denis. "Evolution des effectifs médicaux à l'assistance publique - hôpitaux de Paris - un modèle démographique markovien poissonnien, analyse statistique et résultats." Paris 5, 1997. http://www.theses.fr/1997PA05S011.
Full textDegris, Thomas. "Apprentissage par renforcement dans les processus de décision Markoviens factorisés." Paris 6, 2007. http://www.theses.fr/2007PA066594.
Full textPrasitwattanaseree, Sukon. "Le moment de la transmission du VIH de la mère à l'enfant dépend des durées de traitement par Zidovudine (AZT) de la mère et de l'enfant." Paris 7, 2004. http://www.theses.fr/2004PA077146.
Full textSamson, Vincent. "Approche régularisée pour la détection d'objets ponctuels dans une séquence d'images." Paris 11, 2002. http://www.theses.fr/2002PA112295.
Full textWe address the issue of point target detection in image sequences. Our application deals with optical images of cloud scenes for remote surveillance but this problem is also relevant to other domains such as astronomical or biomedical imaging. Among the different techniques in the literature, clutter removal algorithms identify point objects as pixel anomalies with respect to the background correlated structure. Thus, we propose to solve this problem in a regularized framework by taking into account some a priori information on the two components of the data, cluttered background and isolated point targets. Our goal is to separate them by inverting the observation model. The single frame estimate of the background is defined as the minimizer of a global penalized criterion involving two robust terms. The enhanced target image is the residual output and can be interpreted in a Bayesian framework as the MAP solution. For the sake of computational burden, the criterion is chosen convex and searched by a gradient-type algorithm. We propose an extension of this approach to the simultaneous processing of several frames. It takes into account the apparent translational motion of the background and leads to multiframe estimates of background and targets. Simulation results on realistic images show that the single frame solution overcomes spatial matched filter and in the spatio-temporal case, our method gives improved performance for detecting slow moving objects compared to linear prediction filters. Then, we develop high-resolution models to cape with undersampling, in particular the aliasing of the background. Detection of point objects with random subpixel location is also tackled through the comparison of empirical performances of refined detection statistics
Votsi, Irène. "Evaluation des risques sismiques par des modèles markoviens cachés et semi-markoviens cachés et de l'estimation de la statistique." Thesis, Compiègne, 2013. http://www.theses.fr/2013COMP2058.
Full textThe first chapter describes the definition of the subject under study, the current state of science in this area and the objectives. In the second chapter, continuous-time semi-Markov models are studied and applied in order to contribute to seismic hazard assessment in Northern Aegean Sea (Greece). Expressions for different important indicators of the semi- Markov process are obtained, providing forecasting results about the time, the space and the magnitude of the ensuing strong earthquake. Chapters 3 and 4 describe a first attempt to model earthquake occurrence by means of discrete-time hidden Markov models (HMMs) and hidden semi-Markov models (HSMMs), respectively. A nonparametric estimation method is followed by means of which, insights into features of the earthquake process are provided which are hard to detect otherwise. Important indicators concerning the levels of the stress field are estimated by means of the suggested HMM and HSMM. Chapter 5 includes our main contribution to the theory of stochastic processes, the investigation and the estimation of the discrete-time intensity of the hitting time (DTIHT) for the first time referring to semi-Markov chains (SMCs) and hidden Markov renewal chains (HMRCs). A simple formula is presented for the evaluation of the DTIHT along with its statistical estimator for both SMCs and HMRCs. In addition, the asymptotic properties of the estimators are proved, including strong consistency and asymptotic normality. In chapter 6, a comparison between HMMs and HSMMs in a Markov and a semi-Markov framework is given in order to highlight possible differences in their stochastic behavior partially governed by their transition probability matrices. Basic results are presented in the general case where specific distributions are assumed for sojourn times as well as in the special case concerning the models applied in the previous chapters, where the sojourn time distributions are estimated non-parametrically. The impact of the differences is observed through the calculation of the mean value and the variance of the number of steps that the Markov chain (HMM case) and the EMC (HSMM case) need to make for visiting for the first time a particular state. Finally, Chapter 7 presents concluding remarks, perspectives and future work
Brandejsky, Adrien. "Méthodes numériques pour les processus markoviens déterministes par morceaux." Phd thesis, Bordeaux 1, 2012. http://tel.archives-ouvertes.fr/tel-00733731.
Full textLevernier, Nicolas. "Temps de premier passage de processus non-markoviens." Thesis, Paris 6, 2017. http://www.theses.fr/2017PA066118/document.
Full textThe aim of this thesis is the evaluation of the first-passage time (FPT) of a non-markovian walker over a target. The first part is devoted to the computation of the mean first-passage time (MFPT) for different non-markovien confined processes, for which hidden variables are explicitly known. Our methodology, which adapts an existing formalism, relies on the determination of the distribution of the hidden variables at the instant of FPT. Then, we extend these ideas to the case of general non-markovian confined processes, without introducing the -often unkown- hidden variables. We show that the MFPT is entirely determined by the position of the walker in the future of the FPT. For gaussian walks with stationary increments, this position can be accurately described by a gaussian process, which enable to determine it self-consistently, and thus to find the MFPT. We apply this theory on many examples, in various dimensions. We show moreover that this theory is exact perturbatively around markovian processes. In the third part, we explore the influence of aging properties on the the FPT in confinement, and we predict the dependence of its statistic on geometric parameters. We verify these predictions on many examples. We show in particular that the non-linearity of the MFPT with the confinement is a hallmark of aging. Finally, we study some links between confined and unconfined problems. Our work suggests a promising way to evaluate the persistence exponent of non-markovian gaussian aging processes
Chiquet, Julien. "Modélisation et estimation des processus de dégradation avec application en fiabilité des structures." Phd thesis, Université de Technologie de Compiègne, 2007. http://tel.archives-ouvertes.fr/tel-00165782.
Full textNous étudions la fiabilité du système en considérant la défaillance de la structure lorsque le processus de dégradation dépasse un seuil fixe. Nous obtenons la fiabilité théorique à l'aide de la théorie du renouvellement markovien.
Puis, nous proposons une procédure d'estimation des paramètres des processus aléatoires du système différentiel. Les méthodes d'estimation et les résultats théoriques de la fiabilité, ainsi que les algorithmes de calcul associés, sont validés sur des données simulés.
Notre méthode est appliquée à la modélisation d'un mécanisme réel de dégradation, la propagation des fissures, pour lequel nous disposons d'un jeu de données expérimental.
Rolet, Philippe. "Éléments pour l'Apprentissage et l'Optimisation de Fonctions Chères." Phd thesis, Université Paris Sud - Paris XI, 2010. http://tel.archives-ouvertes.fr/tel-00551865.
Full textFongang, Léon. "Aspects prédictifs des interactions tissus-implant par analyses multi-échelles en imagerie : mise en évidence intraleucocytaires de microtextures décrivant un champ aléatoire markovien." Nice, 1993. http://www.theses.fr/1993NICE4707.
Full textHuman bone and hematologic celle textures are evaluated by the way of fittes biomedical image and signal processsing operators. In the present work, we report a set of algorithmic procedures leading to : 1) a mathematical modeling « cell memory ». That model i related to homogeneity variations associated to the metabolic states which characterize the cell life inside « energetic bands ». The second order statistical means of blood cell microscopic textures agree with those resulting form the markovian random field. 2) A morphological and textural cell classification based on the concept of « field ray-vectorrs » (CRV). Inside that CRV, the micro- and macroscopic aspects of the cell specificities are taken into account fot the contour determination. An indexing method of « Ray Vectors » (RV) leads to a « fictive » (not physically) reorientation of objects, the specific algorithms of rotations and tranlations being excluded. Matched RV give a « shape-signal » (SF) in relation to each object and leads to quantify the degree of similarity between objects. 3) A new concept found upon the morphological multi-scale analysis allowing the quantification of periimplant texture homogeneity and, the predictive evaluations of bone tissue evolutions during the regeneration r after normal or abnormal restructuration which can be more or less altered
Huet, Nolwen. "Inégalités géométriques pour des mesures long-concaves." Toulouse 3, 2009. http://thesesups.ups-tlse.fr/737/.
Full textIn most of this thesis, we study geometric inequalities for some log-concave measures. We give a streamlined semigroup proof of Gaussian Brunn-Minkowski inequality in the case of several sets, with a characterization of the coefficients. Our method also yields semigroup proofs of Brascamp-Lieb inequality and of its reverse form. Then, we show an isoperimetric inequality with a universal constant for isotropic log-concave measures whose density depends only on the radius. This result improves the Cheeger inequality proved by Bobkov. Kannan, Lovász et Simonovits conjectured that any isotropic log-concave measures satisfy a Cheeger inequality with a universal constant. We give new examples for which the conjecture comes true, as uniform measures on convex sets of revolution, and methods to construct other ones. The last part deal with the hypergroup property. It allows the description of all Markov kernels whose eigenvectors are given
Olmedo, Alexandra. "Politiques de stabilisation, fluctuations conjoncturelles, et non-linéarités." Cergy-Pontoise, 2006. http://www.theses.fr/2006CERG0311.
Full textThis thesis is an empirical research on business cycle asymmetries in the conduct and transmission of monetary policy. By business cycle asymmetries, we refer to differences between the expansion and recession phases of the business cycle. Chapter 1 is a survey of the literature existing on the Taylor rule, and provides an answer to the question whether the Taylor rule should be used as a normative or positive tool. Chapter 2 estimates this rule using a markov-switching model pour five OECD economies. Chapter 3 proposes a survey of theoretical and empirical literature regarding the asymmetrical effects (sign, size and cycle) of monetary policy on the economic activity. Chapter 4 introduces models using markov-switching autoregressive vectors (VAR-MS), our definition of the impulse-response function to non-linear shocks, as well as results of Monte-Carlo simulations, evidencing the superiority of our definition compared to the other two existing in the literature. Chapter 5 studies the business cycle asymmetries through the traditional interest rate channel using a VAR-MS estimated for the U. S. Economy. Chapter 6 analyses the business cycle asymmetries through the credit channel using a VAR-MS model estimated for the Euro-zone economy. Results from the last two chapters are in line with the findings of theoretical literature
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.
Full textLiu, Baolong. "Dynamic modeling in sustainable operations and supply chain management." Thesis, Cergy-Pontoise, Ecole supérieure des sciences économiques et commerciales, 2018. http://www.theses.fr/2018ESEC0006.
Full textThis thesis articulates several important issues in sustainable operations and supply chain management not only to provide insights for enhancing the performance of firms but also to appeal to the enterprises to adopt appropriate means for a better environment of our society. The link from firm level to society level is that, to improve the green performance through better operations management efficiency in firms and supply chains, is an indispensable element to ameliorate the environment in our society. Taking China as an example. Since a few years ago (The Straitstimes, 2017; Stanway & Perry, 2018), the government started to spare no effort in resolving the air pollution problems. An important and useful means is to put strict regulations and monitoring the efforts of firms which will face serious fine if certain standards are not met by random inspection. Therefore, firms have to cooperate for the betterment of its profitability and, more importantly, the environmental impacts. Throughout the endeavor, despite the uncertain future situation, the air quality has gradually improved in China (Zheng, 2018). This thesis, in a more general setting, aims to provide important insights to firms so that they are not only able to meet the regulations but genuinely to make contributions to building a better environment for our future generations. Basically, our goal is to obtain deep understanding of the trade-offs with which companies are faced, and to model the problems for seeking possible solutions and helping firms/supply chains to enhance their performance from a theoretical point of view. Then, indirectly, the work will help firms to realize the importance of developing sustainable operations and supply chain management means on our society. The structure of the thesis is organized as follows. Chapter 2 introduces the thesis in French. Chapter 3 is the first essay, Environmental Collaboration and Process Innovation in Supply Chain Management with Coordination. Chapter 4 includes the contents of the second essay, Remanufacturing of Multi-Component Systems with Product Substitution , and the third essay, Joint Dynamic Pricing and Return Quality Strategies Under Demand Cannibalization, is introduced in Chapter 5. Chapter 6 gives the general concluding remarks of the three essays which is followed by the reference list and the appendices
Hachicha, Khalil. "Algorithmes et architectures électroniques pour l'intégration de la détection de mouvement markovienne aux codeurs vidéo MPEG4 / H264." Paris 6, 2005. http://www.theses.fr/2005PA066410.
Full textCauchemez, Simon. "Estimation des paramètres de transmission dans les modèles épidémiques par échantillonnage de Monte Carlo par chaine de Markov." Paris 6, 2005. http://www.theses.fr/2005PA066572.
Full textLimnios, Nikolaos. "Systèmes avec délai de défaillance et processus semi-markoviens." Compiègne, 1991. http://www.theses.fr/1991COMPE091.
Full textGeorgiadis, Stylianos. "Estimation des systèmes semi-markoviens à temps discret avec applications." Thesis, Compiègne, 2013. http://www.theses.fr/2013COMP2112/document.
Full textThe present work concerns the estimation of a discrete-time system whose evolution is governed by a semi-Markov chain (SMC) with finitely many states. We present the invariance principle in a multidimensional form for the semi-Markov kernel (SMK) and some associated measures of the process. Afterwards, we study the nonparametric estimation of the stationary distribution of the SMC, considering two different estimators, and we prove that they hold the same asymptotic behavior. We introduce also the first hitting probability. We propose an estimator and study its asymptotic properties : the strong consistency and the asymptotic normality. On the other hand, we focus on the study of the dependability of semi-Markovsystems. We introduce the interval reliability whose special cases are the reliability and the availability measures and we study the asymptotic properties of a proposed estimator. Moreover, we present a comparison of nonparametric estimation for various reliability measures based on two estimators of the SMK, realizing a unique trajectory and multiple independent observations.Furthermore, this work provides results on the discrete-time semi-Markov case with general state space. We evaluate the average and diffusion approximation of Markov renewal chains. Finally, we are also interested in another class of processes for which we obtain results in the framework of queueing systems. We establish the average approximationfor the Engset model in continuous time and we apply this result to retrial queues
Delorme, Mathieu. "Processus stochastiques et systèmes désordonnés : autour du mouvement Brownien." Thesis, Paris Sciences et Lettres (ComUE), 2016. http://www.theses.fr/2016PSLEE058/document.
Full textIn this thesis, we study stochastic processes appearing in different areas of statistical physics: Firstly, fractional Brownian motion is a generalization of the well-known Brownian motion to include memory. Memory effects appear for example in complex systems and anomalous diffusion, and are difficult to treat analytically, due to the absence of the Markov property. We develop a perturbative expansion around standard Brownian motion to obtain new results for this case. We focus on observables related to extreme-value statistics, with links to mathematical objects: Levy’s arcsine laws and Pickands’ constant. Secondly, the model of elastic interfaces in disordered media is investigated. We consider the case of a Brownian random disorder force. We study avalanches, i.e. the response of the system to a kick, for which several distributions of observables are calculated analytically. To do so, the initial stochastic equation is solved using a deterministic non-linear instanton equation. Avalanche observables are characterized by power-law distributions at small-scale with universal exponents, for which we give new results
Tombuyses, Béatrice. "Modélisation markovienne en fiabilité: réduction des grands systèmes." Doctoral thesis, Universite Libre de Bruxelles, 1994. http://hdl.handle.net/2013/ULB-DIPOT:oai:dipot.ulb.ac.be:2013/212700.
Full textLa première partie de cette thèse concerne Ia modélisation d'installations industrielles et la construction de la matrice de transition. Le but poursuivi est le développement d'un code markovien permettant une description réaliste et aisée du système. Le système est décrit en termes de composants multiétats :pompes, vannes .
La définition d'une série de règles types permet l'introduction de dépendances entre composants. Grâce à la modélisation standardisée du système, un algorithme permettant la construction automatique de la matrice de transition est développé. L'introduction d'opérations de maintenance ou d'information est également présentée.
La seconde partie s'intéresse aux techniques de réduction de la taille de la matrice, afin de rendre possible le traitement de grosses installations. En effet, le nombre d'états croit exponentiellement avec le nombre de composants, ce qui limite habituellement les installations analysables à une dizaine de composants. Les techniques classiques de réduction sont passées en revue :
accessibilité des états,
séparation des groupes de composants indépendants,
symétrie et agrégation exacte des états (cfr Papazoglou). Il faut adapter la notion de symétrie des composants en tenant compte des dépendances pouvant exister entre composants.
Une méthode d'agrégation approchée pour le calcul de la fiabilité et de la disponibilité de groupes de composants à deux états est développée.
La troisième partie de la thèse contient une approche originale pour l'utilisation de la méthode markovienne. Il s'agit du développement d'une technique de réduction basée sur le graphe d'influence des composants. Un graphe d'influence des composants est construit à partir des dépendances existant entre composants. Sur base de ce graphe, un système markovien non homogène est construit, décrivant de manière approchée le comportement du système exact. Les résultats obtenus sur divers exemples sont très bons.
Une quatrième partie de cette thèse s'intéresse aux problèmes numériques liés à l'intégration du système différentiel du problème markovien. Ces problèmes résultent principalement du caractère stiff du système. Différentes méthodes classiques sont implantées pour l'intégration du système différentiel. Elles sont testées sur un exemple type de problème de fiabilité.
Pour finir, on trouve la présentation du code CAMERA dans lequel ont été implantées les différentes techniques présentées ci-dessus.
Doctorat en sciences appliquées
info:eu-repo/semantics/nonPublished
Sun-Hosoya, Lisheng. "Meta-Learning as a Markov Decision Process." Thesis, Université Paris-Saclay (ComUE), 2019. http://www.theses.fr/2019SACLS588/document.
Full textMachine Learning (ML) has enjoyed huge successes in recent years and an ever- growing number of real-world applications rely on it. However, designing promising algorithms for a specific problem still requires huge human effort. Automated Machine Learning (AutoML) aims at taking the human out of the loop and develop machines that generate / recommend good algorithms for a given ML tasks. AutoML is usually treated as an algorithm / hyper-parameter selection problems, existing approaches include Bayesian optimization, evolutionary algorithms as well as reinforcement learning. Among them, auto-sklearn which incorporates meta-learning techniques in their search initialization, ranks consistently well in AutoML challenges. This observation oriented my research to the Meta-Learning domain. This direction led me to develop a novel framework based on Markov Decision Processes (MDP) and reinforcement learning (RL).After a general introduction (Chapter 1), my thesis work starts with an in-depth analysis of the results of the AutoML challenge (Chapter 2). This analysis oriented my work towards meta-learning, leading me first to propose a formulation of AutoML as a recommendation problem, and ultimately to formulate a novel conceptualisation of the problem as a MDP (Chapter 3). In the MDP setting, the problem is brought back to filling up, as quickly and efficiently as possible, a meta-learning matrix S, in which lines correspond to ML tasks and columns to ML algorithms. A matrix element S(i, j) is the performance of algorithm j applied to task i. Searching efficiently for the best values in S allows us to identify quickly algorithms best suited to given tasks. In Chapter 4 the classical hyper-parameter optimization framework (HyperOpt) is first reviewed. In Chapter 5 a first meta-learning approach is introduced along the lines of our paper ActivMetaL that combines active learning and collaborative filtering techniques to predict the missing values in S. Our latest research applies RL to the MDP problem we defined to learn an efficient policy to explore S. We call this approach REVEAL and propose an analogy with a series of toy games to help visualize agents’ strategies to reveal information progressively, e.g. masked areas of images to be classified, or ship positions in a battleship game. This line of research is developed in Chapter 6. The main results of my PhD project are: 1) HP / model selection: I have explored the Freeze-Thaw method and optimized the algorithm to enter the first AutoML challenge, achieving 3rd place in the final round (Chapter 3). 2) ActivMetaL: I have designed a new algorithm for active meta-learning (ActivMetaL) and compared it with other baseline methods on real-world and artificial data. This study demonstrated that ActiveMetaL is generally able to discover the best algorithm faster than baseline methods. 3) REVEAL: I developed a new conceptualization of meta-learning as a Markov Decision Process and put it into the more general framework of REVEAL games. With a master student intern, I developed agents that learns (with reinforcement learning) to predict the next best algorithm to be tried. To develop this agent, we used surrogate toy tasks of REVEAL games. We then applied our methods to AutoML problems. The work presented in my thesis is empirical in nature. Several real world meta-datasets were used in this research. Artificial and semi-artificial meta-datasets are also used in my work. The results indicate that RL is a viable approach to this problem, although much work remains to be done to optimize algorithms to make them scale to larger meta-learning problems
Hofer, Ludovic. "Decision-making algorithms for autonomous robots." Thesis, Bordeaux, 2017. http://www.theses.fr/2017BORD0770/document.
Full textThe autonomy of robots heavily relies on their ability to make decisions based on the information provided by their sensors. In this dissertation, decision-making in robotics is modeled as continuous state and action markov decision process. This choice allows modeling of uncertainty on the results of the actions chosen by the robots. The new learning algorithms proposed in this thesis focus on producing policies which can be used online at a low computational cost. They are applied to real-world problems in the RoboCup context, an international robotic competition held annually. In those problems, humanoid robots have to choose either the direction and power of kicks in order to maximize the probability of scoring a goal or the parameters of a walk engine to move towards a kickable position
Mamane, Salha. "Lois de Wishart sur les cônes convexes." Thesis, Angers, 2017. http://www.theses.fr/2017ANGE0003/document.
Full textIn the framework of Gaussian graphical models governed by a graph G, Wishart distributions can be defined on two alternative restrictions of the cone of symmetric positive definite matrices: the cone PG of symmetric positive definite matrices x satisfying xij=0 for all non-adjacent vertices i and j and its dual cone QG. In this thesis, we provide a harmonious construction of Wishart exponential families in graphical models. Our simple method is based on analysis on convex cones. The focus is on nearest neighbours interactions graphical models, governed by a graph An, which have the advantage of being relatively simple while including all particular cases of interest such as the univariate case, a symmetric cone case, a nonsymmetric homogeneous cone case and an infinite number of non-homogeneous cone cases. The Wishart distributions on QAn are constructed as the exponential family generated from the gamma function on QAn. The Wishart distributions on PAn are then constructed as the Diaconis- Ylvisaker conjugate family for the exponential family of Wishart distributions on QAn. The developed methods are then used to solve the Letac-Massam Conjecture on the set of parameters of type I Wishart distributions on QAn. Finally, we introduce and study exponential families of distributions parametrized by a segment of means with an emphasis on their Fisher information. The focus in on distributions with matrix parameters. The particular cases of Gaussian and Wishart exponential families are further examined
Faddoul, Rafic. "Optimisation de la maintenance des ouvrages de génie civil sur la base des critères de coût et de fiabilité." Nantes, 2010. http://www.theses.fr/2010NANT2108.
Full textWe present in this thesis a Generalized Partially Observable Markov Decision Process GPOMDP which allows for the optimization of complex sequences of decisions during each time period of the planning horizon of the Inspection, Maintenance and Rehabilitation (IM&R) of civil engineering structures. Then, we propose a methodology, using a Lagrangian relaxation technique, for the optimization of the IM&R of multiple structures under the constraints of limited available budgets, creating thus an economical interdependence among the decisions relative to each of the structures of the group. Then, we adapt the multi-structure methodology for the IM&R optimization of series and parallel multi-component systems constrained by minimum reliability limits for the whole system during each stage, thus creating a structural or functional interdependence among the decisions related to each of its components. After that, we propose a methodology which takes into account the epistemic uncertainty veiling the true value of the transition matrices in MDPs. All of the information included in the characterization of the epistemic uncertainty is used. We deal with both cases where the epistemic uncertainty is expressed by means of Bayesian probabilities distributions, and fuzzy sets theory. Finally, we propose an extension of POMDPs in order to take into account, at the beginning of the planning horizon, the possible availability of free information in future time periods. We assume that this information has a Bayesian network structure