Ces travaux de recherche ont porté sur la mise au point et l’optimisation d’un nouveau procédé continu d’élaboration de dispersions stabilisées de particules de polystyrène dans un polyéther polyol. L’objectif final était de parvenir à minimiser le diamètre moyen des particules et à maximiser le taux de solide dans le mélange tout en maintenant sa stabilité.L’étude bibliographique a montré l’importance du choix du stabilisant et des paramètres opératoires du procédé dans la formation de la dispersion. Une étude préliminaire menée en réacteur fermé avec différents stabilisants a conduit à sélectionner un polyol greffé polystyrène. Une étude complète a été réalisée pour mettre au point la synthèse de ce stabilisant et la transposer à l’échelle pilote.Le passage au procédé continu a nécessité la construction d’une unité pilote complète centrée autour d’une extrudeuse bi-vis. Apres optimisation des conditions opératoires, deux paramètres influents ont été mis en évidence : le taux de stabilisant et la concentration du taux de solide. A partir d’un plan d’expériences complet, un modèle simple a été établi rendant compte de l'influence des conditions opératoires sur la qualité des dispersions. L’'optimisation multicritère du procédé et l’utilisation d'outils de génie décisionnel appropriés a permis de déterminer les meilleurs compromis pour l’élaboration des dispersions. / This study dealt with the development and the optimization of a new continuous process for making stable dispersions of polystyrene particles in a polyether polyol. The objective was to maximize the solid mass fraction in the mixture and to minimize the mean particles diameter, by maintaining stable dispersion. The bibliographical research showed the importance of the stabilizer and the operating conditions choices. Preliminary studies carried out in a batch reactor with different stabilizers led to select a polystyrene grafted polyol. A complete study of the synthesis of this stabilizer was carried out and transposed on a pilot scale.For to the continuous process, a complete pilot unit was built around a twin screw extruder. Once the operating conditions were optimized, two influencing parameters were identified (rate of stabilizer and concentration of polystyrene), thus an experimental design was developed. A simple model was created in order to determine the relationships between operating conditions and dispersion’s quality. The use of multicriteria optimization process allowed determining the best compromises for making stable dispersions.

Dans un processus traditionnel de conception de produits mécaniques, l’organisation classique est la conception du produit puis son dimensionnement et enfin son optimisation. Les concepteurs prennent des risques en n’évaluant les performances du produit que dans les dernières phases du processus, et l’étape d’optimisation est bien souvent mise de côté, faute de temps ou à cause du retard pris sur les autres phases. La phase de conception préliminaire est une étape délicate car les décisions doivent y être prises dans un contexte imprécis ou peu de choses sont définies. Il n’existe pas d’outils informatiques adaptés et l’étude de plusieurs concepts multiplie les tâches routinières et les pertes de temps associées. Le concepteur prend donc des risques en réduisant le nombre de concepts développés. Le dimensionnement de produits comporte également un certain nombre de difficultés, comme un positionnement tardif dans le processus, entraînant des étapes de pré et post traitement lourdes ainsi que des simulations longues. L’analyse d’aide à la décision est prometteuse mais elle reste marginale, faute de modèles simplifiés disponibles dans les phases amont du processus de conception.Cette thèse expose la synthèse de nos travaux de recherche portant sur l’optimisation en conception de systèmes mécaniques. L’étude proposée s’intègre dans une démarche d’ingénierie hautement productive de systèmes mécaniques et comporte plusieurs facettes : conception fonctionnelle paramétrique, application d’ingénierie à base de connaissances, outils de simulation numérique et d’optimisation. Nous proposons une méthodologie permettant d’introduire au plus tôt dans le processus de conception, une étape d’optimisation couplée à des simulations d’aide à la décision, en vue d’identifier des architectures optimales inédites, présentant les meilleurs compromis vis-à-vis d’objectifs multiples relatifs au comportement mécanique. L’objectif du travail de recherche est de développer des modèles, des méthodes et des outils dédiés à la génération semi-automatique de modèles géométriques tridimensionnels multiples identifiés comme solutions les plus performantes dans un cadre de conception collaborative de produits mécaniques. Notre méthodologie est ensuite expérimentée sur plusieurs projets de conception de systèmes mécaniques relevant de la mécanique des structures, en analyses statique et dynamique. Il s’agit en particulier de la conception d’organes du véhicule de l’UTBM participant au trophée SIA. / In a traditional process of design of mechanical products, the traditional organization is the product design and the sizing and finally optimization. The designers take risks in failing to assess product performance in the final stages of the process, and the optimization step is often ignored because of time or due to delays on other phases. The preliminary design is a delicate step because decisions must be taken in a little vague or defined. There are no adequate data tools and the study of several concepts multiply routine tasks and wasted time associated. The designer is therefore taking risks by reducing the number of concepts developed. The design of products also includes a number of difficulties, such as positioning late in the process, resulting in steps of pre and post processing of heavy and long simulations. An analysis of decision support is promising but remains marginal due to lack of simplified models available in the early phases of the design process.This thesis describes the synthesis of our research on design optimization of mechanical systems. The proposed study is part of a highly productive process engineering of mechanical systems and multi-faceted: parametric functional design, application engineering knowledge base, simulation tools and optimization. We propose a methodology to introduce at the earliest in the design process, an optimization step coupled with simulations of decision support, to identify optimal architectures novel, the three best compromise vis- -vis multiple objectives related to the mechanical behavior. The objective of the research is to develop models, methods and tools dedicated to the semi-automatic generation of multi-dimensional geometric models identified as most effective solutions in a collaborative design of mechanical products. Our methodology is then tested on several projects of design of mechanical systems within the mechanical structures in static and dynamic analysis. This is especially the design of the vehicle components of the UTBM MLS participant trophy.

We study the problem of minimizing the maximum latency of flows in networks with congestion. We show that this problem is NP-hard, even when all arc latency functions are linear and there is a single source and sink. Still, one can prove that an optimal flow and an equilibrium flow share a desirable property in this situation: all flow-carrying paths have the same length; i.e., these solutions are "fair," which is in general not true for the optimal flow in networks with nonlinear latency functions. In addition, the maximum latency of the Nash equilibrium, which can be computed efficiently, is within a constant factor of that of an optimal solution. That is, the so-called price of anarchy is bounded. In contrast, we present a family of instances that shows that the price of anarchy is unbounded for instances with multiple sources and a single sink, even in networks with linear latencies. Finally, we show that an s-t-flow that is optimal with respect to the average latency objective is near optimal for the maximum latency objective, and it is close to being fair. Conversely, the average latency of a flow minimizing the maximum latency is also within a constant factor of that of a flow minimizing the average latenc

De nos jours, le problème de transport de marchandise occupe une place importante dans la vie économique des sociétés modernes. Le problème de ramassage et de livraison (pick-up and delivery problem) est l’un des problèmes dont une grande partie des chercheurs s’y est intéressée.Il s’agit de déterminer un circuit de plusieurs véhicules, de façon à servir à coût minimal un ensemble de clients et de fournisseurs répartis dans un réseau, satisfaisant certaines contraintes relatives aux véhicules, à leurs capacités et à des précédences entre les nœuds. Les travaux de recherche développés dans cette thèse portent sur le PDPTW (Pickup and Delivery Problem with Time Windows) à plusieurs véhicules (m-PDPTW). Ce dernier a été traité dans les deux cas : statique et dynamique. Nous avons proposé plusieurs approches de résolution du m-PDPTW basées sur les algorithmes génétiques, l’optimisation multicritère et le calcul des bornes inférieures, et ceci pour minimiser un certain nombre de critères comme : le nombre de véhicules utilisés, la somme des retards ou le coût total de transport. Ces approches ont donné de bons résultats, principalement au niveau de la minimisation de la somme des retards où nous avons obtenu, dans plusieurs cas, un retard nul avec un coût de transport tolérable / Nowadays, the transport goods problem occupies an important place in the economic life of modern societies. The PDPTW (Pickup and delivery problem with Time Windows) is one which a large part of researchers was interested. This is an optimization vehicles routing problem which must meet requests for transport between suppliers and customers satisfying precedence and capacity.Researchers developed in this thesis concerns the resolution of the PDPTW with multiple vehicles (m-PDPTW). The latter was treated in two cases: static and dynamic.We have proposed some approaches to solving the m- PDPTW, based on genetic algorithms, multicriteria optimization and the lower bounds, and this to minimize a number of criteria such as: the vehicles number, the total travel cost, and the total tardiness time.Computational results indicate that the proposed approach gives good results with a total tardiness equal to zero with a tolerable cost

[pt] As técnicas de otimização multicritérios, que tiveram sua origem na área econômica, são necessárias na busca de soluções em qualquer ambiente onde existam objetivos competindo entre si. Nestes ambientes pode não existir uma solução ótima para determinado problema e sim soluções de compromisso entre os vários objetivos conflitantes. Uma das muitas áreas onde múltiplos objetivos coexistem é a área de Projetos de Circuitos Eletrônicos, onde desde um compromisso entre ganho e banda passante até compromissos bem mais complexos são freqüentemente encontrados. Para circuitos com alguma complexidade, processos analíticos são virtualmente inviáveis, tornando-se assim indispensável a utilização do computador como instrumento de projeto. Foi desenvolvido, então, um pacote em FORTRAN com aproximadamente 3.000 linhas de instrução para projetar nominalmente circuitos eletrônicos não lineares. Este pacote, respeitando a topologia fornecida para o circuito, atribui valores aos parâmetros de projeto, no caso valores de resistores, capacitores e indutores, de forma a atingir as especificações solicitadas. As opções existentes de projeto que foram implementadas são as seguintes: especificação das tensões nodais DC e/ou polarização de transistores e/ou faixa dinâmica e/ou impedância de entrada e/ou impedância de saída nas freqüências desejadas, além de poder ser escolhida a faixa de variação dos parâmetros de projeto. Estão contidos também, modelos globais (DC/AC) dos dispositivos eletrônicos mais comuns que podem ser solicitados pelo usuário, conforme sua necessidade ao descrever o circuito. Os modelos disponíveis são os seguintes: diodo de junção, transistores bipolares npn e pnp e amplificador operacional com entrada bipolar. / [en] Multiple criterion optimization techniques must be used in environments which competing objetives. In such environments no optimal solution is attained, but only compromise solutions among all the competing objectives. One of these environments is the Designs of Electronic Circuits, whose design compromises generally range from gain and bandwith to much more complex ones. For somewhat complex circuits, analitical processes are virtually unfeasible, so a digital computer has to be used as an instrument for calculus. Thus, a FORTRAN package was written to design nominally non-linear electronic circuits. This package, maintaining the circuit topology unchanged, gives values to resisters, capacitors and inductores so that it may meet the design requirements. The design options provided by the package are: specification of the node voltages and/or total biasing bipolar transistors (Vce - Ic) and/or total power comsumption and/or frequency response and/or input impedance and/or output impedance and/or dynamic range. All these options can be requested limiting the range of variation in the design parameters (resistors, capacitors and inductors) The package still provides global models for semiconductor devides which are: junction diode, bipolar npn/pnp transistors and an operational amplifier macro-model. The design options offered and the general purpose structure of the package provide a favorable environment for the application of the multiple criterion optimization techniques.

Reducing emissions is a very prevailing topic also in aviation industry. Besides technological improvements it is necessary to also adjust procedures and operations. We present a tool that is able to optimize flight trajectories not only regarding economical factors but also ecological ones. Due to the utilization of a flight performance model and a detailed engine model, it is possible to determine an aircraft’s emission quantities during all phases of a flight. Besides those like CO2, NOx, CO, Black Carbon, etc. we are also considering contrail formation and their influence on the global warming. By transforming emissions into monetary values we can find a trade off between multiple criteria. After describing the employed models and software architecture we present some use cases where we successfully applied our toolchain.

Geometric errors may compromise the quality of radiation therapy treatments. Optimization methods that account for errors can reduce their effects. The first paper of this thesis introduces minimax optimization to account for systematic range and setup errors in intensity-modulated proton therapy. The minimax method optimizes the worst case outcome of the errors within a given set. It is applied to three patient cases and shown to yield improved target coverage robustness and healthy structure sparing compared to conventional methods using margins, uniform beam doses, and density override. Information about the uncertainties enables the optimization to counterbalance the effects of errors. In the second paper, random setup errors of uncertain distribution---in addition to the systematic range and setup errors---are considered in a framework that enables scaling between expected value and minimax optimization. Experiments on a phantom show that the best and mean case tradeoffs between target coverage and critical structure sparing are similar between the methods of the framework, but that the worst case tradeoff improves with conservativeness. Minimax optimization only considers the worst case errors. When the planning criteria cannot be fulfilled for all errors, this may have an adverse effect on the plan quality. The third paper introduces a method for such cases that modifies the set of considered errors to maximize the probability of satisfying the planning criteria. For two cases treated with intensity-modulated photon and proton therapy, the method increased the number of satisfied criteria substantially. Grasping for a little less sometimes yields better plans. In the fourth paper, the theory for multicriteria optimization is extended to incorporate minimax optimization. Minimax optimization is shown to better exploit spatial information than objective-wise worst case optimization, which has previously been used for robust multicriteria optimization. The fifth and sixth papers introduce methods for improving treatment plans: one for deliverable Pareto surface navigation, which improves upon the Pareto set representations of previous methods; and one that minimizes healthy structure doses while constraining the doses of all structures not to deteriorate compared to a reference plan, thereby improving upon plans that have been reached with too weak planning goals. / <p>QC 20130516</p>

This thesis deals with the mapping and the scheduling of workflows. In this context, we consider unreliable platforms, with processors subject to failures. In a first part, we consider a particular model of streaming applications : the filtering services. In this context, we aim at minimizing period and latency. We first neglect communication costs. In this model, we study scheduling problems on homogeneous and heterogeneous platforms. Then, the impact of communication costs on scheduling problems of a filtering application is studied. Finally, we consider the scheduling problem of such an application on a chain of processors. The theoretical complexity of any variant of this problem is proved. This filtering property can model the reliability of processors. The results of some computations are successfully computed, and some other ones are lost. We consider the more frequent failure types : transient failures. We aim efficient and reliable schedules. The complexity of many variants of this problem is proved. Two heuristics are proposed and compared using using simulations. Even if transient failures are the most common failures in classical grids, some particular type of platform are more concerned by other type of problems. Desktop grids are especially unstable. In this context, we want to execute iterative applications. All tasks are executed, then a synchronization occurs, and so on. Two variants of this problem are considered : applicationsof independent tasks, and applications where all tasks need to be executed at same speed. In both cases, the problem is first theoretically studied, then heuristics are proposed and compared using simulations.

This thesis addresses two important classes of optimization : multiobjective optimization and bilevel optimization. The investigation concerns their solution methods, applications, and possible links between them. First of all, we develop a procedure for solving Multiple Objective Linear Programming Problems (MOLPP). The method is based on a new characterization of efficient faces. It exploits the connectedness property of the set of ideal tableaux associated to degenerated points in the case of degeneracy. We also develop an approach for solving Bilevel Linear Programming Problems (BLPP). It is based on the result that an optimal solution of the BLPP is reachable at an extreme point of the underlying region. Consequently, we develop a pivoting technique to find the global optimal solution on an expanded tableau that represents the data of the BLPP. The solutions obtained by our algorithm on some problems available in the literature show that these problems were until now wrongly solved. Some applications of these two areas of optimization problems are explored. An application of multicriteria optimization techniques for finding an optimal planning for the distribution of electrical energy in Cameroon is provided. Similary, a bilevel optimization model that could permit to protect any economic sector where local initiatives are threatened is proposed. Finally, the relationship between the two classes of optimization is investigated. We first look at the conditions that guarantee that the optimal solution of a given BPP is Pareto optimal for both upper and lower level objective functions. We then introduce a new relation that establishes a link between MOLPP and BLPP. Moreover, we show that, to solve a BPP, it is possible to solve two artificial M0PPs. In addition, we explore Bilevel Multiobjective Programming Problem (BMPP), a case of BPP where each decision maker (DM) has more than one objective function. Given a MPP, we show how to construct two artificial M0PPs such that any point that is efficient for both problems is also efficient for the BMPP. For the linear case specially, we introduce an artificial MOLPP such that its resolution can permit to generate the whole feasible set of the leader DM. Based on this result and depending on whether the leader can evaluate or not his preferences for his different objective functions, two approaches for obtaining efficient solutions are presented

[pt] No Brasil, o planejamento da operação energética do Sistema Interligado Nacional - SIN é realizado atualmente por meio de uma cadeia de modelos matemáticos concebidos para otimizar o planejamento segundo o critério de minimização do valor esperado do custo total de operação. No entanto, desde a crise ocorrida no Setor Elétrico Brasileiro entre os anos de 2001 e 2002, cujo ápice ocorreu no racionamento de energia, houve uma intensificação na busca por métodos de otimização que permitam a consideração explicita de critérios adicionais na otimização, tal como a segurança operativa. Neste contexto, este trabalho propõe uma modelagem utilizando algoritmos genéticos que permite a consideração de múltiplos objetivos no processo de otimização sem que a representação física do sistema e de suas incertezas se- jam comprometidas. A abordagem multicritério para o problema possibilita que diversos indicadores, dentre os quais destaca-se o risco anual de déficit, que atualmente são apenas resultados da otimização, se tornem controláveis a partir de sua consideração diretamente no processo de otimização. A modelagem proposta foi computacionalmente implementada na linguagem C# utilizando a biblioteca GAcom desenvolvida pelo ICA/PUC-Rio. O desempenho da metodologia proposta foi avaliado por meio de estudos de casos aplicados ao SIN. Os resultados obtidos, assim como as vantagens observadas ao se utilizar a otimização multicritério, são discutidos ao longo do texto. / [en] In Brazil, the planning of the energy operation of the National Interconnected Power System is currently done through a chain of mathematical models designed to optimise the planning according to criterion of minimisation of the expected value of the total operation`s cost. However, since the 2001-2002 energy supply crisis, there was an intensification in the search for methods of optimization allowing explicit consideration of additional criteria, such as the operative security. In this context, this work proposes a modelling using genetic algorithms that makes possible the consideration of multiple objectives in the optimisation process without compromising the physical representation of the system and its uncertainties. A multicriteria approach to the problem allows that various indicators, like, for instance, annual deficit, which currently are only results of the optimisation, become controllable from its consideration in the optimisation process. The modelling proposal was computationally implemented in language C# using the GAcom library developed by the ICA/PUC-Rio. The performance of the proposed methodology was evaluated through potential National Inter- connected Power System case studies. The results, as well as the benefits seen when using the multicriteria optimisation, are discussed throughout the text.