Spelling suggestions: "subject:" loptimisation"" "subject:" doptimisation""
171 |
Optimisation spatio-temporelle d’efforts de recherche pour cibles manoeuvrantes et intelligentes / Spatio-temporal optimisation of search efforts for smart and reactive moving targetsChouchane, Mathieu 17 October 2013 (has links)
Dans cette thèse, nous cherchons à répondre à une problématique formulée par la DGA Techniques navales pour surveiller une zone stratégique : planifier le déploiement spatial et temporel optimal d’un ensemble de capteurs de façon à maximiser les chances de détecter une cible mobile et intelligente. La cible est dite intelligente car elle est capable de détecter sous certaines conditions les menaces que représentent les capteurs et ainsi de réagir en adaptant son comportement. Les déploiements générés pouvant aussi avoir un coût élevé nous devons tenir compte de ce critère lorsque nous résolvons notre problématique. Il est important de noter que la résolution d’un problème de ce type requiert, selon les besoins, l’application d’une méthode d’optimisation mono-objectif voire multiobjectif. Jusqu’à présent, les travaux existants n’abordent pas la question du coût des déploiements proposés. De plus la plupart d’entre eux ne se concentrent que sur un seul aspect à la fois. Enfin, pour des raisons algorithmiques, les contraintes sont généralement discrétisées.Dans une première partie, nous présentons un algorithme qui permet de déterminer le déploiement spatio-temporel de capteurs le plus efficace sans tenir compte de son coût. Cette méthode est une application à l’optimisation de la méthode multiniveau généralisée.Dans la seconde partie, nous montrons d’abord que l’utilisation de la somme pondérée des deux critères permet d’obtenir des solutions sans augmenter le temps de calcul. Pour notre seconde approche, nous nous inspirons des algorithmes évolutionnaires d’optimisation multiobjectif et adaptons la méthode multiniveau généralisée à l’optimisation multiobjectif. / In this work, we propose a solution to a problem issued by the DGA Techniques navales in order to survey a strategic area: determining the optimal spatio-temporal deployment of sensors that will maximize the detection probability of a mobile and smart target. The target is said to be smart because it is capable of detecting the threat of the sensors under certain conditions and then of adapting its behaviour to avoid it. The cost of a deployment is known to be very expensive and therefore it has to be taken into account. It is important to note that the wide spectrum of applications within this field of research also reflects the need for a highly complex theoretical framework based on stochastic mono or multi-objective optimisation. Until now, none of the existing works have dealt with the cost of the deployments. Moreover, the majority only treat one type of constraint at a time. Current works mostly rely on operational research algorithms which commonly model the constraints in both discrete space and time.In the first part, we present an algorithm which computes the most efficient spatio-temporal deployment of sensors, but without taking its cost into account. This optimisation method is based on an application of the generalised splitting method.In the second part, we first use a linear combination of the two criteria. For our second approach, we use the evolutionary multiobjective optimisation framework to adapt the generalised splitting method to multiobjective optimisation. Finally, we compare our results with the results of the NSGA-II algorithm.
|
172 |
Analýza sortimentu Macromex Czech s.r.o. / Optimisation of the product lines of the trading companyMaskaliova, Maryia January 2010 (has links)
Analyse of the product lines of the trading company based on the next methods: analyse of revenues, analyse of profitability, analyse od effectivity, analyse of the status of a product in the line. Optimisation of the product lines based on the results of analyse
|
173 |
Conception d'un cadre d'optimisation de fonctions d'énergies : application au traitement d'images / New framework design for optimizing energy functions : application to image processingKouzana, Amira 14 December 2018 (has links)
Nous proposons une nouvelle formulation de minimisation de fonctions d’énergies pour la traitement de la vision sur toute la segmentation d'image. Le problème est modélisé comme étant un jeu stratégique non coopératif, et le processus d'optimisation est interprété comme étant la recherche de l'équilibre de nash. Ce problème reste un problème combinatoire sous cette forme d'où nous avons opté à le résoudre en utilisant un algorithme de Séparation-Évaluation. Pour illustrer la performance de la nouvelle approche, nous l'avons appliqué sur des fonctions de régularisation convexe ainsi que non convexe / We propose a new formulation of the energy minimisation paradigm for image segmentation. The segmentation problem is modeled as a non-cooperative strategic game, and the optimization process is interpreted as the search of a Nash equilibrium. The problem is expressed as a combinatorial problem, for which an efficient Branch and Bound algorithm is proposed to solve the problem exactly. To illustrate the performance of the proposed framework, it is applied on convex regularization model, as well as a non-convex regularized segmentation models
|
174 |
Modélisation et optimisation non convexe basées sur la programmation DC et DCA pour la résolution de certaines classes des problèmes en fouille de données et cryptologie / The non-convex modeling and optimization based on the DC programming and DCA for the resolution of certain classes of problems in Data Mining and cryptologyLe, Hoai Minh 24 October 2007 (has links)
Cette thèse est consacrée à la modélisation et l'optimisation non convexe basées sur la programmation DC et DCA pour certaines classes de problèmes issus de deux domaines importants : le Data Mining et la Cryptologie. Il s'agit des problèmes d'optimisation non convexe de très grande dimension pour lesquels la recherche des bonnes méthodes de résolution est toujours d'actualité. Notre travail s'appuie principalement sur la programmation DC et DCA. Cette démarche est motivée par la robustesse et la performance de la programmation DC et DCA, leur adaptation aux structures des problèmes traités et leur capacité de résoudre des problèmes de grande dimension. La thèse est divisée en trois parties. Dans la première partie intitulée Méthodologie nous présentons des outils théoriques servant des références aux autres. Le premier chapitre concerne la programmation DC et DCA tandis que le deuxième porte sur les algorithmes génétiques. Dans la deuxième partie nous développons la programmation DC et DCA pour la résolution de deux classes de problèmes en Data Mining. Dans le chapitre quatre, nous considérons le modèle de la classification floue FCM et développons la programmation DC et DCA pour sa résolution. Plusieurs formulations DC correspondants aux différentes décompositions DC sont proposées. Notre travail en classification hiérarchique (chapitre cinq) est motivé par une de ses applications intéressante et très importantes, à savoir la communication multicast. C'est un problème non convexe, non différentiable de très grande dimension pour lequel nous avons reformulé sous la forme des trois programmes DC différents et développé les DCA correspondants. La troisième partie porte sur la Cryptologie. Le premier concerne la construction des fonctions booléennes équilibrées de haut degré de non-linéarité - un des problèmes cruciaux en Cryptographie. Plusieurs versions de combinaison de deux approches - DCA et les algorithmes génétiques (AG) sont étudiées dans le but d'exploiter simultanément l'efficacité de chaque approche. Le deuxième travail concerne des techniques de cryptanalyse d'un schéma d'identification basé sur les deux problèmes ''Perceptron'' (PP) et ''Perceptron Permuté'' (PPP). Nous proposons une méthode de résolution des deux problèmes PP et PPP par DCA et une méthode de coupes dans le dernier chapitre / This thesis is dedicated to non-convex modeling and the optimization based on the DC programming and DCA for certain classes of problems of two important domains : the Data Mining and the Cryptology. They are non-convex optimization problems of very large dimensions for which the research of good solution methods is always of actuality. Our work is based mainly on the DC programming and DCA that have been successfully applied in various fields of applied sciences, including machine learning. It is motivated and justified by the robustness and the good performance of DC programming and DCA in comparison with the existing methods. This thesis is devised in three parties. The first part, entitling Methodology, serves as a reference for other chapters. The first chapter concerns the programming of DC and DCA while the second chapter describes the genetic algorithms. In the second part, we develop the DC and DCA programming to solve two classes of problems in Data Mining. In the chapter four, we take consideration into the model of classification FCM and develop the programming DC and DCA for their resolution. Many formulations DC in correspondence to different decompositions DC are proposed. Our work in hierarchic classification (chapter 5) is motivated by one of its interesting and very important applications, known as muliticast communication. It's a non-convex, non differentiable, non-convex problem in a very big dimension with which we have reformulated in the forms of 3 different DC programs and developed the DCA relative. The 3rd part focuses on the Cryptology. The 1st chapter is the construction of stable boonlean functions with high degree of non-linearity - one of the crucial problems of Cryptography. Many versions of combination of 2 approaches, DCA and Genetic Algorithms (GA) are studied in the purpose of exploiting simultaneously the efficacy of each approach. The secondrd work is about the techinics of cryptanalyse of a identification scheme based on two problems Perceptron (PP) and Perceptron Permuted. We propose a method of resolving two problems PP and PPA by DCA and a cutting plan method in the last chapter
|
175 |
Analyse et modélisation de formes optimales / Analysis and modeling of optimal shapesDurus, Ioana-Geanina 28 November 2008 (has links)
L'objet de ce travail de thèse est l'étude théorique et numérique des quelques problèmes relevants de l'analyse et de la modélisation de forme. Les problèmes considérés sont issus d'applications modernes comme la modélisation de décollement de membrane par mouvements minimisants, des inégalités isopérimétriques et de traitement d'image. Analyse et optimisation de formes des valeurs propres de l'opérateur de conductivité. Dans un ouvert de mesure finie on considère le Laplacian avec conditions aux bord de type conductivité, i.e. Constantes localement ou globalement, avec constantes libres. Cet opérateur intervient dans le processus de détection de défauts par mesures au bord et a fait l'objet d'une première analyse par Greco et Lucia dans le cadre globalement constant. Nous étudions des propriétés qualitatives des valeurs propres en relation avec la géométrie, des inégalités isopérimétriques générales par réarrangement et/ou [gamma]-convergence, et nous implémentons un algorithme génétique pour déterminer les formes minimisantes pour les valeurs propres d'ordre petit, à mesure constante, La génération des formes est basée sur les niveaux des séries de Fourier tronquées, contrôlées par les coefficients. Simulation numérique du décollement d'un membrane. On étudie le modèle de décollement d'une membrane adhésive proposé par Bucur, Buttazzo et Lux, dans le cadre des mouvements minimisantes quasi-statiques. Suivant la dissipation de l'énergie, le décollement consiste en une évolution de domaines ou une évolution de mesures. Mon travail a consisté dans la simulation associées à des stratégies évolutionnaires en relation avec la dérivée de forme, méthodes des lignes de niveaux et la dérivation des mesures. Etude numérique du problème de localisation optimale. Nous nous intéressons à l'analyse numérique des différents problèmes de localisation optimale tels que le problème de chois optimal des pixels d'appui pour l'interpolation d'image ou la minimisation de la compliance (cadre linéaire ou non linéaire). Différentes méthodes sont utilisées, tels que le gradient topologique, les algorithmes génétiques et le calcul asymptotique par [gamma]-convergence / We are interested in a few problems which can be viewed in the framework of shape optimization like : the conductivity eigenvalue problem. In an open set of finite measure we considers the Laplacian with conductivity boundary conditions, i.e., constants locally or globally with free constants. The operator intervenes in the process of the defect identication of a material by electrostatic boundary measurements and was the subject of a first properties of the eigenvalues in relation with the geometry, general isoperimetric inequalities by rearrangement and/or [gamma]-convergence, and we implement a genetic with constant measure. The generation of the forms is based on the levels of truncated Fourier series, controlled by the coefficeints. Numerical implementation for debonding membranes. We study the model of quasistatic evolution of an adhesive membrane subjected to a debonding force proposed by Bucur, Buttazzo and Lux, within the framwork of the minimizing movements. According to the dissipation of energy the debonding process consisted in the numerical simulation of the debonding process in the two situations. Methods of descent are associated with the evolution strategies in relation to the shape derivative ; level set method and the mesure derivation. Numerical results for a few problem of optimal location. We are interested in the numeraical analysis of the various problems of optimal location such as the problem of best alternative of the pixels of support for the image interpolation or minimization of compliance (framework linear and not linear). Various methods are used, such as the topological gradient, the genetic algorithms and asymptotic calculation by [gamma]-convergence
|
176 |
Contribution à l'optimisation des plans d'exploitation et de maintenance, selon une approche basée sur le pronostic : application au domaine naval / Contribution to the optimization of operating plans and maintenance, using an approach based on prognosis : application to the naval domainSchutz, Jérémie 03 December 2009 (has links)
Dans le domaine naval, lorsqu'un navire quitte le quai, il s'engage à réaliser un ensemble de missions durant un horizon de temps déterminé. L'ensemble des missions à accomplir ainsi que l'ordonnancement de celles-ci constitue le plan d'exploitation. Comme un navire peut réaliser différents types de missions (diplomatique, océanographique, etc.), la dégradation du système dépend des conditions opérationnelles et environnementales. L'objectif consiste donc à choisir parmi toutes les missions proposées, dont la durée cumulée dépasse l'horizon de temps alloué, celles à réaliser ainsi que l'ordonnancement. Aussi, afin de minimiser les coûts de maintenance, différentes stratégies de maintenance seront étudiées telles que les politiques systématique, sporadique périodique et séquentielle. De manière générale, les politiques de maintenance sont constituées d'actions correctives minimales et préventives imparfaites. Compte tenu des hypothèses de travail (horizon de temps fini, loi de défaillance dynamique) l'optimisation de ces politiques de maintenance sera réalisée à l'aide de procédures numériques ou de méta-heuristiques. L'optimisation des plans d'exploitation étant résolue à l'aide méta-heuristiques, l'optimisation conjointe des deux plans peut résulter de l'intégration de deux méta-heuristiques. Le travail de recherche a donc pour objectif le développement d'un outil d'aide à la décision basé sur le pronostic. Il doit permettre de générer les plans d'exploitation et de maintenance optimaux compte tenu de l'historique des missions réalisées, du retour d'expérience d'autres navires mais aussi des missions génériques possibles / The present research work aims to develop a decision-making tool based on prognosis. Indeed, in the naval domain, when a navy ship leaves the dock, it should experience a set of missions within a finite time horizon. The business plan consist to perform as well as to schedule the overall missions. As a navy ship can achieve different type of missions (diplomatic, oceanographic, etc.), the system degradations varie according to both operational and environmental conditions. Therefore, the business plan consist on choosing among all possible missions those whose the cumulative duration does not exceed the finite time horizon. Such missions should be then scheduled and performed. On the other hand, our work deals with different maintenance policies such as systematic, sporadic, periodic and sequential. Generally, the maintenance actions are of two types, namely minimal corrective and inperfect preventive. Based on some assumptions (finite time horizon, dynamic failure law), the optimization of maintenance policy will be carried out using exact numerical procedures or meta-heuristic. To solve the joint optimization of business and maintenance plans, we use the integration of two types of meta-heuristic. The obtained solution should define the optimal business and maintenance plans given the history of missions performed, the experiences of other navy ships but also of all possible generic missions
|
177 |
Algorithms for decision makingRiseth, Asbjørn Nilsen January 2018 (has links)
We investigate algorithms for different steps in the decision making process, focusing on systems where we are uncertain about the outcomes but can quantify how probable they are using random variables. Any decision one makes in such a situation leads to a distribution of outcomes and requires a way to evaluate a decision. The standard approach is to marginalise the distribution of outcomes into a single number that tries in some way to summarise the value of each decision. After selecting a marginalisation approach, mathematicians and decision makers focus their analysis on the marginalised value but ignore the distribution. We argue that we should also be investigating the implications of the chosen mathematical approach for the whole distribution of outcomes. We illustrate the effect different mathematical formulations have on the distribution with one-stage and sequential decision problems. We show that different ways to marginalise the distributions can result in very similar decisions but each way has a different complexity and computational cost. It is often computationally intractable to approximate optimal decisions to high precision and much research goes into developing algorithms that are suboptimal in the marginalised sense, but work within the computational budget available. If the performance of these algorithms is evaluated they are mainly judged based on the marginalised values, however, comparing the performance using the full distribution provides interesting information: We provide numerical examples from dynamic pricing applications where the suboptimal algorithm results in higher profit than the optimal algorithm in more than half of the realisations, which is paid for with a more significant underperformance in the remaining realisations. All the problems discussed in this thesis lead to continuous optimisation problems. We develop a new algorithm that can be used on top of existing optimisation algorithms to reduce the cost of approximating solutions. The algorithm is tested on a range of optimisation problems and is shown to be competitive with existing methods.
|
178 |
Modelling and optimisation of bistable composite laminatesBetts, David January 2012 (has links)
Asymmetric composite laminates can have a bistable response to loading. The potentially large structural deformations which can be achieved during snap-through from one stable state to another with small and removable energy input make them of interest for a wide range of engineering applications. After 30 years of research effort the shapes and response to applied loads of laminates of general layup can be quantitatively predicted. With attention switching to the incorporation of bistable laminates for practical applications, tools for the design and optimisation of actuated bistable devices are desirable. This thesis describes the analytical and experimental studies undertaken to develop novel modelling and optimisation techniques for the design of actuated asymmetric bistable laminates. These structures are investigated for practical application to morphing structures and the developing technology of piezoelectric energy harvesting. Existing analytical models are limited by the need for a numerical solver to determine stable laminate shapes. As the problem has multiple equilibria, convergence to the desired solution cannot be guaranteed and multiple initial guesses are required to identify all possible solutions. The approach developed in this work allows the efficient and reliable prediction of the stable shapes of laminates with off-axis ply orient at ions in a closed form manner. This model is validated against experimental data and finite element predictions, with an extensive sensitivity study presented to demonstrate the effect of uncertainty and imperfections in the laminate composition. This closed-form solution enables detailed optimisation studies to tailor the design of bistable devices for a range of applications. The first study considers tailoring of the directional stiffness properties of bistable laminates to provide resistance to externally applied loads while allowing low energy actuation. The optimisation formulation is constrained to guarantee bistability and to ensure a useful level of deformation. It is demonstrated that 'cross-symmetric' layups can provide stiffness in an arbitrary loading direction which is five times greater than in a chosen actuation direction.
|
179 |
The application of multi-attribute optimisation as a systems engineering tool in an automotive CAE environmentSutton, Paul January 2012 (has links)
Multi-Attribute Optimisation (MAO) is proposed as a tool for delivering high value products within the systems engineering approach taken in the automotive industry. This work focuses on MAO methods that use Computer Aided Engineering (CAE) analyses to build a metamodel of system behaviour. A review of the literature and current Jaguar Land Rover optimisation methods showed that the number of samples required to build a metamodel could be estimated using the number of input variables. The application of these estimation methods to a concept airbox design showed that this guidance may not be sufficient to fully capture the complexity of system behaviour in the metamodelling method. The use of the number of input variables and their ranges are proposed as a new approach to the scaling of sample sizes. As a corollary to the issue of the sample size required for accurate metamodelling, the sample required to estimate the error was also examined. This found that the estimation of the global error by additional samples may be impractical in the industrial context. CAE is an important input to the MAO process and must balance the efficiency and accuracy of the model to be suitable for application in the optimisation process. Accurate prediction of automotive attributes may require the use of new CAE techniques such as multi-physics methods. For this, the fluid structure interaction assessment of the durability of internal components in the fuel tank due to slosh was examined. However, application of the StarCD-Abaqus Direct couple and Abaqus Combined Eularian Lagrangian was unsuitable for this fuel slosh application. Further work would be required to assess the suitability of other multi-physics methods in an MAO architecture. Application of the MAO method to an automotive airbox shows the potential for improving both product design and lead time.
|
180 |
Engineering chemistry : integrated control strategies and Internet-enabled tools for chemical synthesisFitzpatrick, Daniel Ewert January 2017 (has links)
The use of enabling technologies and continuous methods to enhance chemical synthesis is a vibrant area of research, gaining increasing attention from laboratories in academia and industry. Yet many prominent synthesis procedures have not changed for decades and require significant manual intervention from bench chemists, which may lead to the waste of both human and material resources. The research described herein details how chemistry has been bridged with engineering to address this issue in a world increasingly focussed on sustainability and efficiency. This thesis is divided into four chapters. The first describes the development of a novel Internet-based process control system which is applied to automate a cycling catalytic process. The capability of the system to conduct multi-dimensional self-optimisation processes is shown, where it is integrated with an on-line mass spectrometer and inline infrared spectrometer to drive optimisation against customisable multicomponent evaluation functions. Chapter 2 details the successful synthesis of the anti-cancer drug candidate AZ82. The control system is applied to assist with a number of these steps, including facilitating the integration between batch and flow processes on a single reactor platform. A new distillation unit to assist with downstream solvent switching is also described. In Chapter 3, the cloud-based nature of the control system is exploited by moving it to servers residing in Tokyo. The ability of the system to accelerate the drug development process is highlighted with the autonomous self-optimisation and synthesis of four active pharmaceutical ingredient targets: tramadol, lidocaine, bupropion and isoniazid. In the case of bupropion, the system maintained steady-state operation of a telescoped two-step process for an extended period. A researcher in Los Angeles was able to initiate and monitor all processes, via Japan, in real-time as they occurred in our laboratory in Cambridge, UK. Finally, Chapter 4 describes the development of a new parallel column supercritical fluid chromatography (SFC) unit that is capable of separating a multicomponent product stream continuously exiting a flow reactor. The versatility of the SFC unit is showcased with the telescoped synthesis of isoniazid in which all stages, including the SFC process itself, are managed by the control system without researcher intervention.
|
Page generated in 0.2748 seconds