No description available.

No description available.

No description available.

Dans un contexte économique fortement concurrentiel, la maintenance des équipements de productions constitue un enjeu socio-économique décisif pour les entreprises. De même, les enjeux de maintenance ont connu de fortes évolutions, le secteur est en pleine mutation. Face à cela, les attentes, les besoins en matière de coût, de délai, de qualité et de respect de l’environnement changent. Suite à ce constat, on a souhaité développer dans cette thèse de nouvelles stratégies intégrées de maintenance, de production, de gestion des stocks de pièces de rechange en tenant compte de différentes contraintes opérationnelles. Parmi ces contraintes, on peut citer la variation de la cadence de production des systèmes et son impact sur la dégradation du système de production ainsi que sur le plan de maintenance et de gestion de pièces de rechanges. On s’est intéressé aussi à l’origine et au type des pièces de rechange (neuves ou usagées) utilisées pour assurer les activités de maintenance. En plus de l’influence technico-économique de l’origine des pièces de rechange utilisées sur le système de production, on a également évalué l’impact environnemental de ces derniers en considérant des données de l’ADEME. Formellement, L’objectif de cette thèse est de contribuer par des approches formelles permettant une optimisation des cadences de production couplée à des politiques de maintenance et intégrant des stratégies efficientes dans la gestion des pièces de rechange. Des méthodes analytiques et techniques d’optimisation, appuyées par des modèles de simulations selon la difficulté de résolutions ont été développées afin de réaliser les objectifs de cette thèse. Une étude de cas industrielle ainsi que des exemples numériques sont présentés pour valider les résultats théoriques obtenus / In a highly competitive economic environment, maintenance of production facilities is a crucial socio-economic challenge for companies. Similarly, maintenance issues experienced significant changes, the sector is changing. Faced with this, the expectations, the needs of cost, time, quality and environmental change. Following this, we wanted to develop in this thesis new integrated strategies for maintenance, production, and inventory management of spare parts taking into account different operational constraints. These constraints include the change in the rate of production systems and its impact on the deterioration of the production system and on the maintenance plan and spare parts management. It was also interested in the origin and type of spare parts (new or used) used to ensure maintenance activities. In addition to the technical and economic influence of the origin of spare parts used in the production system, we also evaluated the environmental impact of the latter considering ADEME data. Formally, The aim of this thesis is to contribute by formal approaches to optimization of production rates coupled with maintenance policies and integrating efficient strategies in the management of spare parts. Analytical methods and optimization techniques, supported by simulation models according to resolutions of difficulty have been developed to achieve the objectives of this thesis. An industrial case studies and numerical examples are presented to validate theoretical results

During thesis brings second-order cybernetics into design research, in the context of the Internet of Things (IoT) and ‘smart’ homes. My main proposition is to question and critically analyse the embedded epistemology in IoT technology in relation to human centred activities. I examine how human lives are represented within the quantified approaches inherent in current notions of ‘smart’ technology, derived from Artificial Intelligence (AI), and characterise this as the Algorithmic Paradigm. I explore questions of how complex, lived, human experience is oversimplified in the IoT. By adopting an epistemology derived from second-order cybernetics — acknowledging the importance of the observer — combined with my ‘IdIoT Proposition’, a way of ‘slowing down’ research on a fast-paced topic, I explore designing reflectively. The IdIoT is a methodological framework characterised by the process of slowing down and asking ‘What are we busy doing?’ in order to become aware of algorithmic oversimplifications. This methodological approach provides self- awareness and self-reflection on ‘the way of knowing the world’ to the researcher and to the participants, in the context of the Algorithmic Paradigm applied in IoT. Through a series of practice-based projects, I use the figure of the ‘SMART’ fridge to examine the implications of the Algorithmic Paradigm in the ‘smart’ home. The consideration that ‘smartness’ is relational is investigated in Becoming Your ‘SMART’ Fridge, in which I position myself as the algorithm behind a ‘smart’ fridge, using quantitative and qualitative data to make sense and ‘nonsense’ outcomes, and exploring householders’ interpretations. In the ‘SMART’ Fridge Session, I developed scripted dialogues characterised by active, reflective users, and assigned roles in which the ‘smartness’ of the algorithms is explored via professional performances and fictitious roles taken on by members of the public. The findings reveal the value of second-order cybernetics, acknowledging an unpredictable observer and embracing ‘smart’ as relational in interaction with IoT technology. They suggest that a shift in perspective is required to create more meaningful interactions with devices in the ‘smart’ home, questioning the current technological path, challenging the dominant epistemology and proposing alternatives. My methodological approach demonstrates how design research and 1 second-order considerations can work together, asking novel questions to inform disciplines with an interest in the IoT, both from a design perspective and in terms of broader implications for society. The work has value for design, HCI, Critical Algorithm Studies, and for technical developers involved in the creation of IoT systems.

In this doctoral thesis we consider various property testing problems for structured distributions. A distribution is said to be structured if it belongs to a certain class which can be simply described in approximation terms. Such distributions often arise in practice, e.g. log-concave distributions, easily approximated by polynomials (see [Bir87a]), often appear in econometric research. For structured distributions, testing a property often requires far less samples than for general unrestricted distributions. In this thesis we prove that this is indeed the case for several distance-related properties. Namely, we give explicit sub-linear time algorithms for L1 and L2 distance testing between two structured distributions for the cases when either one or both of them are available as a “black box”. We also prove that the given algorithms have the best possible asymptotic complexity by proving matching lower bounds in the form of explicit problem instances (albeit constructed using randomized techniques) demanding at least a specified amount of data to be tested successfully. As the main numerical result, we prove that testing that total variation distance to an explicitly given distribution is at least e requires O(√k/e²) samples, where k is an approximation parameter, dependent on the class of distribution being tested and independent of the support size. Testing that the total variation distance between two “black box” distributions is at least e requires O(k⁴/⁵e⁶/⁵). In some cases, when k ~ n, this result may be worse than using an unrestricted testing algorithm (which requires O( n²/3/e² ) samples where n is the domain size). To address this issue, we develop a third algorithm, which requires O(k²/³e⁴/³ log⁴/³(n/k) log log(n/k)) and serves as a bridge between the cases of small and large domain sizes.

Cette thèse est consacrée à l'étude de l’optimisation combinatoire multi-objective sous incertitudes. Plus particulièrement, nous abordons les problèmes multi-objectifs contenant des données floues qui sont exprimées par des nombres triangulaires floues. Pour faire face à ce type de problèmes, notre idée principale est d’étendre les concepts multi-objectifs classiques au contexte flou. Nous proposons, dans un premier temps, une nouvelle approche Pareto entre des objectifs flous (i.e. vecteurs des nombres triangulaires flous). Ensuite, nous étendons des méta-heuristiques basées sur Pareto afin de converger vers des solutions optimales floues. L’approche proposée est illustrée sur un problème bi-objectif de routage de véhicules avec des demandes floues. Dans le deuxième volet de ce travail, nous abordons l’aspect de robustesse dans le contexte multi-objectif flou en proposant une nouvelle méthodologie d’évaluation de robustesse des solutions. Finalement, les résultats expérimentaux sur des benchmarks flous du problème de routage de véhicules prouvent l’efficacité et la fiabilité de notre approche. / This thesis is devoted to the study of multi-objective combinatorial optimization under uncertainty. In particular, we address multi-objective problems with fuzzy data, in which fuzziness is expressed by fuzzy triangular numbers. To handle such problems, our main idea is to extend the classical multi-objective concepts to fuzzy context. To handle such problems, we proposed a new Pareto approach between fuzzy-valued objectives (i.e. vectors of triangular fuzzy numbers). Then, an extension of Pareto-based metaheuristics is suggested as resolution methods. The proposed approach is thereafter illustrated on a bi-objective vehicle routing problem with fuzzy demands. At the second stage, we address robustness aspect in the multi-objective fuzzy context by proposing a new methodology of robustness evaluation of solutions. Finally, the experimental results on fuzzy benchmarks of vehicle routing problem prove the effectiveness and reliability of our approach.

This thesis is concerned with deriving planning algorithms for robot manipulators. Manipulation has two effects, the robot has a physical effect on the object, and it also acquires information about the object. This thesis presents algorithms that treat both problems. First, I present an extension of the well-known piano mover’s problem where a robot pushing an object must plan its movements as well as those of the object. This requires simultaneous planning in the joint space of the robot and the configuration space of the object, in contrast to the original problem which only requires planning in the latter space. The effects of a robot action on the object configuration are determined by the non-invertible rigid body mechanics. To solve this a two-level planner is presented that coordinates planning in each space. Second, I consider planning under uncertainty and in particular planning for information effects. I consider the case where a robot has to reach and grasp an object under pose uncertainty caused by shape incompleteness. The main novel outcome is to enable tactile information gain planning for a dexterous, highdegree of freedom manipulator with non- Gaussian pose uncertainty. The method is demonstrated in trials with both simulated and real robots.

Diese Forschungsarbeit beschreibt alle Aspekte der Entwicklung eines neuartigen, autonomen Quadrokopters, genannt AQopterI8, zur Innenraumerkundung. Dank seiner einzigartigen modularen Komposition von Soft- und Hardware ist der AQopterI8 in der Lage auch unter widrigen Umweltbedingungen autonom zu agieren und unterschiedliche Anforderungen zu erfüllen. Die Arbeit behandelt sowohl theoretische Fragestellungen unter dem Schwerpunkt der einfachen Realisierbarkeit als auch Aspekte der praktischen Umsetzung, womit sie Themen aus den Gebieten Signalverarbeitung, Regelungstechnik, Elektrotechnik, Modellbau, Robotik und Informatik behandelt. Kernaspekt der Arbeit sind Lösungen zur Autonomie, Hinderniserkennung und Kollisionsvermeidung. Das System verwendet IMUs (Inertial Measurement Unit, inertiale Messeinheit) zur Orientierungsbestimmung und Lageregelung und kann unterschiedliche Sensormodelle automatisch detektieren. Ultraschall-, Infrarot- und Luftdrucksensoren in Kombination mit der IMU werden zur Höhenbestimmung und Höhenregelung eingesetzt. Darüber hinaus werden bildgebende Sensoren (Videokamera, PMD), ein Laser-Scanner sowie Ultraschall- und Infrarotsensoren zur Hindernis-erkennung und Kollisionsvermeidung (Abstandsregelung) verwendet. Mit Hilfe optischer Sensoren kann der Quadrokopter basierend auf Prinzipien der Bildverarbeitung Objekte erkennen sowie seine Position im Raum bestimmen. Die genannten Subsysteme im Zusammenspiel erlauben es dem AQopterI8 ein Objekt in einem unbekannten Raum autonom, d.h. völlig ohne jedes externe Hilfsmittel, zu suchen und dessen Position auf einer Karte anzugeben. Das System kann Kollisionen mit Wänden vermeiden und Personen autonom ausweichen. Dabei verwendet der AQopterI8 Hardware, die deutlich günstiger und Dank der Redundanz gleichzeitig erheblich verlässlicher ist als vergleichbare Mono-Sensor-Systeme (z.B. Kamera- oder Laser-Scanner-basierte Systeme). Neben dem Zweck als Forschungsarbeit (Dissertation) dient die vorliegende Arbeit auch als Dokumentation des Gesamtprojektes AQopterI8, dessen Ziel die Erforschung und Entwicklung neuartiger autonomer Quadrokopter zur Innenraumerkundung ist. Darüber hinaus wird das System zum Zweck der Lehre und Forschung an der Universität Würzburg, der Fachhochschule Brandenburg sowie der Fachhochschule Würzburg-Schweinfurt eingesetzt. Darunter fallen Laborübungen und 31 vom Autor dieser Arbeit betreute studentische Bachelor- und Masterarbeiten. Das Projekt wurde ausgezeichnet vom Universitätsbund und der IHK Würzburg-Mainfranken mit dem Universitätsförderpreis der Mainfränkischen Wirtschaft und wird gefördert unter den Bezeichnungen „Lebensretter mit Propellern“ und „Rettungshelfer mit Propellern“. Außerdem wurde die Arbeit für den Gips-Schüle-Preis nominiert. Absicht dieser Projekte ist die Entwicklung einer Rettungsdrohne. In den Medien Zeitung, Fernsehen und Radio wurde über den AQopterI8 schon mehrfach berichtet. Die Evaluierung zeigt, dass das System in der Lage ist, voll autonom in Innenräumen zu fliegen, Kollisionen mit Objekten zu vermeiden (Abstandsregelung), eine Suche durchzuführen, Objekte zu erkennen, zu lokalisieren und zu zählen. Da nur wenige Forschungsarbeiten diesen Grad an Autonomie erreichen, gleichzeitig aber keine Arbeit die gestellten Anforderungen vergleichbar erfüllt, erweitert die Arbeit den Stand der Forschung. / The following scientific work describes entirely the development of an innovative autonomous quadrotor, called AQopterI8, for indoor exploration and object search describing solutions in the fields of signal processing, control theory, software and hardware. The main topics are autonomy, collision avoidance and obstacle detection. The AQopterI8 uses an IMU (inertial measurement unit) for attitude control and together with ultrasonic, infrared and pressure sensors for height control. Furthermore ultrasonic, infrared, stereo vision, pmd (photonic mixing device) and laser sensors enable the quadrotor to fly autonomously, detect obstacles and avoid collisions (distance control). All ressources are on-Board and the quadrotor requires no external system such as GPS or OTS (optical tracking system) cameras. The evaluation shows, that the AQopterI8 can control its distance to moving obstacles such as persons, can fly through corridors and search, localize and count autonmously target objects. The work was marked with the university promotion prize from the chamber of commerce (IHK Mainfranken) and has been many times in the media such as newspaper, radio and television.

This thesis is concerned with the evaluation of properties of early design models, the control structure selection and the representation and properties of composite system models. The work is motivated by the need to introduce a Systems and Control Theory based framework for early design stages of the overall system design of engineering processes, and in particular chemical processes, such as process synthesis and control structure selection. The overall spirit of the work is that engineering design is an evolutionary process, the different stages of which shape the structure of the resulting system models and precondition the potential for design at the next stage. The work identifies a number of key problems in the overall design, which are of a generic, systemic character, and then deals with working out solutions for such problems. The results contribute in the development of a framework for systems integration using as criteria and tools, those provided by Systems and Control Theory. The work aims to provide a control theoretic dimension to the rules and practices currently used in the specific application areas. The thesis contributes in the development of a Systems and Control conceptual and tools framework for integrated design of engineering processes by providing results in the following areas: - Specification of a number of generic problems in the field of integrated design and identification of relevant control theoretic concepts and tools. - Study of Model Orientation for linear implicit state-space models and definition of classes of oriented realisations. - Development of solutions to two problems of Structural Identification for uncertain early process models related to infinite zero structure and McMillan degree. - Development of a generic representation of composite systems that allows the study of transition from the aggregate to composite system properties, as a generalised Control Design and characterisation of some key system properties. -Specification of a framework for integrated Control Structure Selection and development of software for many approaches of the "interaction analysis" indicators. The results contribute in the area of development of the systems and control ideas for the problems of systems integration and early design. The work emphasises the strong links between Modelling, Selection of Control Structures and Control Structural methodologies.