Motivated by the lack of models that can bring a general preparation formula for developing competitive intelligence in Trade Show, the authors researched in a theoretical database to develop a model that can bring a general vision for a company that wants to start to organize the personnel to gather competitive intelligence in trade show events.          Furthermore, was discovered that not many firms do actually realize the opportunities they have to acquire competitive intelligence in trade shows, that way the model was developed to facilitate the process.         The discovering of the research pointed Socialization as well as relationship building and strengthening were the main channels to acquire information. In addition, Trade Shows create a rich environment where most of the stakeholders are present and also willing to share knowledge and information, creating an even more favorable place to develop this kind of intelligence.         The model presented takes in consideration the company’s stakeholder, and how to allocate personnel in booths and in extensive research around the area, it also classifies the visitors and attendees in different groups, facilitating the organizing process to understand and explore more easily and efficiently the Trade Show. The models also suggest approaches to each group in order to don’t invest resources in an ineffective way.

Neste trabalho analisou-se o problema da manutenção, no longo prazo, dos produtores de leite na atividade. Foram analisados estabelecimentos nos estados de Rondônia, Tocantins e Rio de Janeiro. O que se pode notar é um significativo grupo que se mantém por longos anos produzindo pouco, sendo que possui capacidade para aumentar a produção. As restrições que impedem este grupo de se especializar, estão ligadas à uma série de imperfeições inerentes ao mercado e, que acabam tornando dos produtores, principalmente os pequenos, vítimas deste sistema. O objetivo principal do trabalho foi verificar a existência de economias de escala entre os produtores de leite dos estados de Rondônia, Tocantins e Rio de Janeiro. Para isso foram estimadas cinco funções custo com diferentes níveis de restrições. A função que apresentou melhor aderência aos dados foi a de custo translog. Os fatores de produção considerados foram capital, terra, trabalho e os dispêndios diretos. A análise econômica mostra a dificuldade de sobrevivência dos estabelecimentos no longo prazo. Isto ocorre porque a relação capital imobilizado/produção é muito alta. Este capital também se mostra concentrado em investimentos que indicam baixo nível tecnológico, como por exemplo, no fator terra e em animais com baixa aptidão para a atividade leiteira. O investimento em máquinas e equipamentos foi baixo nos três estados analisados, não sendo superior a 5% do total de capital imobilizado. Em relação à composição do rebanho nos estados, nota-se, sobretudo entre os pequenos produtores, a presença de muitos animais nas categorias de novilha, bezerro e machos em recria e engorda. A mão-deobra utilizada na maioria dos estabelecimentos era de origem familiar. Apenas no estado do Rio de Janeiro percebe-se maior quantidade de mão-de-obra contratada. Quanto à taxa de retorno da atividade, nenhum dos três estados apresentou taxas maiores que 6%, mesmo quando o custo da terra era excluído do cálculo. Ressalta-se que o estado do Rio de Janeiro foi o que apresentou maiores taxas de retorno. Nota-se uma relação direta entre os níveis de produtividade e as taxas de retorno. Os resultados da regressão revelam que a grande maioria dos produtores da amostra está na faixa de economias de escala, sendo que apenas 3,4% destes estão na faixa de deseconomias de escala. O ponto de custo médio mínimo foi obtido na faixa de 178 mil litros por ano, ou seja, média diária de 487 litros por dia. Outro grupo composto por 10% da amostra está mais próximo do ponto de custo médio mínimo e apresenta produção diária entre 183 e 487 litros por dia. E finalmente, o último grupo que aparece em situação menos favorável, apresentando uma produção inferior a 180 litros/dia. Destaca-se que estes podem reduzir de forma significativa os custos se aumentarem a produção, já que se encontram na faixa mais acentuada da curva de custo médio de longo prazo e podem crescer bastante na atividade usufruindo das economias de escala. / This study aimed to analize the business maintenance issue faced by milk producers, in the long term. It was analized farms in the states of Rondônia, Tocantins and Rio de Janeiro. What can be noticed is a group that has maintained itself for years with little production, whereas it is able to increase it. The restrictions that prevent the group from specializing, are linked to a series of imperfections inherent to the market, which end up turning producers, especially the small ones, into victms of this system. The main goal of this study was to verify the existence of economies of scale among milk producers in the states of Rondônia, Tocantins and Rio de Janeiro. In order to do so, it was estimated five cost functions at different levels of restrictions. The function that presented better adequacy to the data was that of cost translog. The production factors taken into consideration were: capital, land, work and direct expenses. The economic analysis shows the difficulties to survive faced by farmers in the long term. This happens because the relation immobilized capital/production is very high. This capital is also found in investments which indicate low technological level, as an example, in the land factor and in animals with low milk production. The investment in machinery and equipment was low in the three states studied, without reaching 5% of the whole immobilized capital. As for the compositon of livestocks in the states, it is noticed, especially among small producers, the presence of many animals in the categories of heifers, calves, males for breeding and fattening. The workforce used in most farms was of family origin. Only in the state of Rio de Janeiro it was noticed a higher number of hired workforce. As for the return rate of the activity, none of the three states presented rates higher than 6%, even when the value of the land was excluded from the figure. It is pointed out that the state of Rio de Janeiro showed higher return rates. It is observed a direct relation between the productivity levels of and the return rates. The regression results reveal that the vast majority of producers in this case study operate in economies of scale, being that only 3.4% of them do not. The point of minimum average cost was obtained at about 178 thousand liters per year, that is, an average of 487 liters per day. Another group formed by 10% of the case study is closer to the point of minimum average cost and presents a daily production between 183 and 487 liters. And finally, the last group which presents itself in a less favorable situation with a production lower than 180 liters per day. It is highlighted that these producers can significantly reduce their costs if they increase their production, once they are found in the most accentuated part of the average cost cuve in the long term and can greatly develop in the activity by using the economies of scale.

abstract: 中国商品期货市场经历30年发展，已初备协调资源分配、对冲经营风险的功能。但受产业自身和期货市场发展的制约，各期货品种市场有效性参差不齐。随着我国经济从增量阶段过渡到存量阶段，期货作为企业的价格管理和风险控制工具的重要性日益凸显，因此研究我国商品期货市场有效性具有非常好的现实意义。 本文开创性的从期货的基本功能——资源配置的角度出发，提出有效市场是指其期货价格能够对本行业社会资源起到合理的调配作用的市场。在内容安排上，本文首先总结了现有国际成熟期货品种的特点并找出能够反映期货对资源配置能力的四个指标假说，分别为期现回归性、利润波动性、库存波动性以及现金流变化，然后通过数学模型证明指标数据和品种成熟度的关联，最后应用该套指标对我国商品市场有效性进行检验。数学方法上，本文先采用Bai-Perron内生多重结构突变模型对时间序列进行突变点检验，然后对断点时间序列分别进行多元回归，并在剔除季节性和周期性后，通过平稳性检验、ARCH效应检验结果来确定相应的Garch模型，并用Garch模型来描述时间序列的波动性。 通过数学验证，我们认为期现回归性、利润波动性、库存波动性以及现金流变化这四个指标可以作为反映期货成熟度的检验指标，用该套方法对国内部分活跃品种检验后发现大连豆粕期货已经具备成熟品种的特征，本文认为豆粕期货市场是有效的；PTA、玉米淀粉期货的四个检验指标在近年来表现出时间序列优化的特点，但因时间较短尚不稳定，可以认为是接近成熟的品种；而螺纹钢和铝期货在多数指标上表现不佳，表明他们对社会资源配置能力较差，因此本文认为螺纹钢和铝期货市场是活跃但非有效的。通过进一步分析，本文认为品种的期现回归性差是制约其资源配置能力发挥的关键因素，而交易标的不明确、 仓单制作难度大、产业参与度低以及期货设计中的其他限制因素又是导致期现回归性差的重要原因。 / Dissertation/Thesis / Doctoral Dissertation Business Administration 2019

Neste trabalho analisou-se o problema da manutenção, no longo prazo, dos produtores de leite na atividade. Foram analisados estabelecimentos nos estados de Rondônia, Tocantins e Rio de Janeiro. O que se pode notar é um significativo grupo que se mantém por longos anos produzindo pouco, sendo que possui capacidade para aumentar a produção. As restrições que impedem este grupo de se especializar, estão ligadas à uma série de imperfeições inerentes ao mercado e, que acabam tornando dos produtores, principalmente os pequenos, vítimas deste sistema. O objetivo principal do trabalho foi verificar a existência de economias de escala entre os produtores de leite dos estados de Rondônia, Tocantins e Rio de Janeiro. Para isso foram estimadas cinco funções custo com diferentes níveis de restrições. A função que apresentou melhor aderência aos dados foi a de custo translog. Os fatores de produção considerados foram capital, terra, trabalho e os dispêndios diretos. A análise econômica mostra a dificuldade de sobrevivência dos estabelecimentos no longo prazo. Isto ocorre porque a relação capital imobilizado/produção é muito alta. Este capital também se mostra concentrado em investimentos que indicam baixo nível tecnológico, como por exemplo, no fator terra e em animais com baixa aptidão para a atividade leiteira. O investimento em máquinas e equipamentos foi baixo nos três estados analisados, não sendo superior a 5% do total de capital imobilizado. Em relação à composição do rebanho nos estados, nota-se, sobretudo entre os pequenos produtores, a presença de muitos animais nas categorias de novilha, bezerro e machos em recria e engorda. A mão-deobra utilizada na maioria dos estabelecimentos era de origem familiar. Apenas no estado do Rio de Janeiro percebe-se maior quantidade de mão-de-obra contratada. Quanto à taxa de retorno da atividade, nenhum dos três estados apresentou taxas maiores que 6%, mesmo quando o custo da terra era excluído do cálculo. Ressalta-se que o estado do Rio de Janeiro foi o que apresentou maiores taxas de retorno. Nota-se uma relação direta entre os níveis de produtividade e as taxas de retorno. Os resultados da regressão revelam que a grande maioria dos produtores da amostra está na faixa de economias de escala, sendo que apenas 3,4% destes estão na faixa de deseconomias de escala. O ponto de custo médio mínimo foi obtido na faixa de 178 mil litros por ano, ou seja, média diária de 487 litros por dia. Outro grupo composto por 10% da amostra está mais próximo do ponto de custo médio mínimo e apresenta produção diária entre 183 e 487 litros por dia. E finalmente, o último grupo que aparece em situação menos favorável, apresentando uma produção inferior a 180 litros/dia. Destaca-se que estes podem reduzir de forma significativa os custos se aumentarem a produção, já que se encontram na faixa mais acentuada da curva de custo médio de longo prazo e podem crescer bastante na atividade usufruindo das economias de escala. / This study aimed to analize the business maintenance issue faced by milk producers, in the long term. It was analized farms in the states of Rondônia, Tocantins and Rio de Janeiro. What can be noticed is a group that has maintained itself for years with little production, whereas it is able to increase it. The restrictions that prevent the group from specializing, are linked to a series of imperfections inherent to the market, which end up turning producers, especially the small ones, into victms of this system. The main goal of this study was to verify the existence of economies of scale among milk producers in the states of Rondônia, Tocantins and Rio de Janeiro. In order to do so, it was estimated five cost functions at different levels of restrictions. The function that presented better adequacy to the data was that of cost translog. The production factors taken into consideration were: capital, land, work and direct expenses. The economic analysis shows the difficulties to survive faced by farmers in the long term. This happens because the relation immobilized capital/production is very high. This capital is also found in investments which indicate low technological level, as an example, in the land factor and in animals with low milk production. The investment in machinery and equipment was low in the three states studied, without reaching 5% of the whole immobilized capital. As for the compositon of livestocks in the states, it is noticed, especially among small producers, the presence of many animals in the categories of heifers, calves, males for breeding and fattening. The workforce used in most farms was of family origin. Only in the state of Rio de Janeiro it was noticed a higher number of hired workforce. As for the return rate of the activity, none of the three states presented rates higher than 6%, even when the value of the land was excluded from the figure. It is pointed out that the state of Rio de Janeiro showed higher return rates. It is observed a direct relation between the productivity levels of and the return rates. The regression results reveal that the vast majority of producers in this case study operate in economies of scale, being that only 3.4% of them do not. The point of minimum average cost was obtained at about 178 thousand liters per year, that is, an average of 487 liters per day. Another group formed by 10% of the case study is closer to the point of minimum average cost and presents a daily production between 183 and 487 liters. And finally, the last group which presents itself in a less favorable situation with a production lower than 180 liters per day. It is highlighted that these producers can significantly reduce their costs if they increase their production, once they are found in the most accentuated part of the average cost cuve in the long term and can greatly develop in the activity by using the economies of scale.

The limited overcrowded radio frequency spectrum compelled researchers to ex plore higher frequency ranges for wireless transmission. In recent decades, visible light communications (VLC) have gained lots of research attention thanks to the abundant bandwidth and the existing lighting infrastructure they offer. Throughout this dissertation, we study the downlink of multi-user VLC systems with the aim of operation and architecture enhancement. In this context, we accommodate the chal lenges imposed by the visible light nature such as illumination requirements and mod ulation constraints. On the operation optimization front, we investigate three VLC setups: indoor single cell, outdoor energy harvesting enabled single cell, and indoor energy harvesting enabled multi-cell VLC systems. We formulate, and provide low complexity solutions to, resource allocation problems for each setup while accounting for scenario-tailored system objectives and quality of service requirements. For the first setup, the temporal average illumination is maintained fixed while maximizing the system SE and dynamic time-division-multiple-access is employed to serve users in an interference free setup. As for the second setup, owing to the favored joint lighting and SE maximization, we solve a multi-objective optimization problem accounting for both objectives. We found that the severity of the illumination - communications tradeoff increases as the available system power budget decreases or the minimum rate requirements get tighter. In the third setup, transmitters average currents and receivers fields of view tuning strategies are developed to maximize both spectral ef ficiency and energy harvesting objectives in an interference limited scenario, where spatial illumination uniformity is required. It is found that receivers fields of view tuning is substantial to performance enhancement in dense deployments. On the architecture optimization front, we propose two intelligent reflecting surfaces-aided VLC systems and derive their power density distribution in the receiver plane. In addition, we prove their power concentration capability and quantify their relative gain with respect to one another and with respect to the reflector-free VLC systems enjoying direct line of sight. Finally, we study the channel impulse response of the proposed reflecting systems and quantify the incurred delay spread through exact ex pression, simplified bounds and asymptotic expressions when the number of reflecting elements grows unboundedly.

Dans les systèmes de communications sans fils, l'allocation de ressources reste toujours un défi considérable afin de satisfaire les demandes des utilisateurs et de fournir de bonnes performances avec une perpétuelle demande en applications gourmandes en ressources. Les techniques multiporteuses essentiellement les techniques dérivant de l'OFDM sont généralement utilisées pour transmettre les données dans des sous-porteuses orthogonales. De plus, de nouvelles stratégies de relayage sont proposées pour améliorer les performances en bordures de cellules. Plusieurs types de relais peuvent être sollicités comme les relais fixes faisant partie de l'infrastructure du système ou les relais mobiles qui ne nécessitent pas un coût additionnel de déploiement.Dans cette thèse, nous étudions principalement l'allocation des ressources pour le sens montant d'un système cellulaire OFDMA assurant les exigences de qualité de service et l'équité entre les utilisateurs. Les algorithmes d'allocation de ressources les plus utilisés sont présentés et un nouvel algorithme se basant sur l'équité proportionnelle pondérée (WPF) est proposé afin d'approcher les bornes supérieures de débit et d'équité. L'algorithme WPF propose un poids variable par utilisateur permettant d'allouer un nombre plus élevé de sous-porteuses au centre de la cellule qu'en bordure tout en gardant une bonne équité entre les utilisateurs. Nous établissons une étude théorique afin de comparer l'algorithme proposé à l'algorithme classique d'équité proportionnelle (PF). Nous étendons ensuite l'algorithme WPF à un système multi-cellulaire où l'interférence inter-cellulaire (ICI) dégrade les performances du système. Enfin, nous étudions les stratégies d'annulation de l'ICI et proposons une nouvelle méthode pour réduire l'ICI en se basant sur la coopération entre les stations de base (BSs) et sur les indicateurs d'interférence. Nous proposons un nouvel indicateur d'interférence (EII) à valeurs entières échangé par les BSs pour indiquer les niveaux d'interférence des sous-porteuses. En prenant en compte les valeurs de EII échangées, chaque BS alloue dynamiquement les sous-porteuses de manière à éviter de fortes valeurs d'ICI.Dans un deuxième temps, nous étudions la communication coopérative en utilisant des relais mobiles. Pour cela, de simples utilisateurs mobiles ayant des positions avantageuses peuvent relayer d'autres utilisateurs en bordure de cellule en plus de transmettre leurs propres données. Un relai utilisant le protocole DF multiplexe ainsi ses propres données aux données relayées avant de transmettre à la BS. L'allocation de ressource est formulée sous forme d'un problème d'optimisation dont le but est de minimiser la puissance totale d'émission du système tout en assurant un débit cible par utilisateur. Dans un premier temps, nous proposons une méthode de sélection des relais comme phase d'initialisation et offrons une heuristique itérative pour optimiser l'allocation de puissance et des blocs de ressources radio (RBs). Dans un second temps, nous traitons la sélection des relais comme une variable d'optimisation additionnelle. Pour la résolution, la décomposition de Lagrange et la méthode duale sont utilisées et le problème global est divisé en sous problèmes résolus de manière itérative afin d'approcher la solution optimale. Enfin, nous étendons ce modèle coopératif à un modèle à antennes multiples (MIMO) afin d'étudier l'influence des antennes multiples sur la puissance totale de transmission. Les paramètres à optimiser sont la sélection des relais, l'allocation des puissances et l'allocation des RBs. Afin d'allouer la puissance sur les antennes d'un utilisateur, nous avons étudié la répartition égale des puissance (EPA) et le beamforming. Les expressions théoriques correspondantes sont établies et les résultats de simulation sont présentés pour comparer le modèle avec EPA et le modèle avec beamforming au modèle non coopératif. / In wireless systems, resource allocation is still an important challenge to satisfy user requirements and to ensure good system performances with always greedy data applications. Multicarrier techniques especially the Orthogonal Frequency Division Multiplexing (OFDM) techniques are generally used to carry data into orthogonal subcarriers. Furthermore, relaying strategies are used to enhance cell edge performances. Many types of relays can be investigated as fix relays being part of the network infrastructure or mobile relays without additional deployment cost.In this thesis, we mainly consider the resource allocation for an uplink Orthogonal Frequency Division Multiple Access (OFDMA) system for a cellular system model ensuring Quality of Service (QoS) requirements and fairness between users. The most used resource allocation algorithms are presented and a novel Weighted Proportional Fair (WPF) algorithm is proposed to approach upper bounds of both throughput and fairness. The WPF algorithm considers user weights to allocate more subcarriers in the cell center than in the cell edge keeping sufficient fairness between users. We establish a theoretical analysis to compare the behavior of the proposed WPF algorithm to the classical Proportional Fair (PF) algorithm. Then, we extend this WPF algorithm to a multi-cell system model where the Inter-Cell Interference (ICI) limits the system performance. Moreover, we study ICI mitigation strategies and propose a novel method to reduce the ICI based on Base Station (BS) cooperation and interference indicators. We propose the Enhanced Interference Indicator (EII) with integer values to be exchanged by the BSs indicating interference levels for the subcarriers. Function of these communicated EII values, each BS allocates dynamically subcarriers in order to reduce the ICI. Our contributions in the multi-cell system model are the WPF and the EII.Moreover, we investigate in this dissertation the cooperative communication using mobile relays and propose multiple contributions. For this, simple mobile users with advantageous positions can relay cell edge users to carry data to the BS in addition to their own data. A Decode and Forward (DF) relay multiplex then its own data and relayed data before transmitting to the BS. The resource allocation is formulated as an optimization problem aiming to minimize the system transmit power and respecting a required target data rate per user constraint. In a first time, we propose an initialization method for the paring step to associate source-relay pairs and propose an iterative heuristic to optimize both power and Resource Blocks (RB) allocations. In a second time, we consider the relay selection as an optimization variable in addition of power and RB allocations. For resolution, Lagrangian decomposition and Dual method are used and the global problem is divided into subproblems iterativelly resolved to approach the optimal solution. Finally, we extend this cooperative system model to a Multiple Input Multiple Output (MIMO) system model to study the influence of multiple antennas on the system transmit power. The features to optimize are relay selection, power and RBs allocation. Moreover, to allocate power in the different antennas for each user, both Equal Power Allocation (EPA) and beamforming are studied. Theoretical expressions are established and simulations results are presented to compare EPA, beamforming and non-cooperative system.

No description available.

abstract: In the recent past, Iraq was considered relatively rich considering its water resources compared to its surroundings. Currently, the magnitude of water resource shortages in Iraq represents an important factor in the stability of the country and in protecting sustained economic development. The need for a practical, applicable, and sustainable river basin management for the Tigris and Euphrates Rivers in Iraq is essential. Applicable water resources allocation scenarios are important to minimize the potential future water crises in connection with water quality and quantity. The allocation of the available fresh water resources in addition to reclaimed water to different users in a sustainable manner is of the urgent necessities to maintain good water quantity and quality. In this dissertation, predictive water allocation optimization models were developed which can be used to easily identify good alternatives for water management that can then be discussed, debated, adjusted, and simulated in greater detail. This study provides guidance for decision makers in Iraq for potential future conditions, where water supplies are reduced, and demonstrates how it is feasible to adopt an efficient water allocation strategy with flexibility in providing equitable water resource allocation considering alternative resource. Using reclaimed water will help in reducing the potential negative environmental impacts of treated or/and partially treated wastewater discharges while increasing the potential uses of reclaimed water for agriculture and other applications. Using reclaimed water for irrigation is logical and efficient to enhance the economy of farmers and the environment while providing a diversity of crops, especially since most of Iraq’s built or under construction wastewater treatment plants are located in or adjacent to agricultural lands. Adopting an optimization modelling approach can assist decision makers, ensuring their decisions will benefit the economy by incorporating global experiences to control water allocations in Iraq especially considering diminished water supplies. / Dissertation/Thesis / Doctoral Dissertation Civil, Environmental and Sustainable Engineering 2019

This dissertation solves an optimization problem in the area of scheduling large-scale petroleum development projects under several resources constraints. The dissertation focuses on the application of a metaheuristic search Genetic Algorithm (GA) in solving the problem. The GA is a global search method inspired by natural evolution. The method is widely applied to solve complex and sizable problems that are difficult to solve using exact optimization methods. A classical resource allocation problem in operations research known under Knapsack Problems (KP) is considered for the formulation of the problem. Motivation of the present work was initiated by certain petroleum development scheduling problem in which large-scale investment projects are to be selected subject to a number of resources constraints in several periods. The constraints may occur from limitations in various resources such as capital budgets, operating budgets, and drilling rigs. The model also accounts for a number of assumptions and business rules encountered in the application that motivated this work. The model uses an economic performance objective to maximize the sum of Net Present Value (NPV) of selected projects over a planning horizon subject to constraints involving discrete time dependent variables. Computational experiments of 30 projects illustrate the performance of the model. The application example is only illustrative of the model and does not reveal real data. A Greedy algorithm was first utilized to construct an initial estimate of the objective function. GA was implemented to improve the solution and investigate resources constraints and their effect on the assets value. The timing and order of investment decisions under constraints have the prominent effect on the economic performance of the assets. The application of an integrated optimization model provides means to maximize the financial value of the assets, efficiently allocate limited resources and to analyze more scheduling alternatives in less time.

La localisation active et passive par un réseau de capteurs distribués est un problème rencontré dans différents domaines d’application. En localisation active, telle que la localisation par radar MIMO (Multiple Input Multiple Output), les émetteurs transmettent des signaux qui sont réfléchis par les cibles visées, puis captés par les antennes réceptrices, alors qu’en localisation passive, les capteurs reçoivent des signaux transmis par les cibles elles-mêmes. L’objectif de cette thèse est d’étudier différentes techniques d’optimisation pour la localisation active et passive de haute précision. Dans la première partie de la thèse, on s’intéresse à la localisation active, où de multiples émetteurs illuminent les cibles depuis différentes directions. Les signaux peuvent être émis avec des puissances ou des largeurs de bande différentes. Ces différentes ressources, par nature en général fortement limitées, sont souvent, par défaut, réparties de façon uniforme entre les différents émetteurs. Or, la précision de la localisation dépend de la position des émetteurs, ainsi que des paramètres (les gains notamment) des différents canaux existant entre émetteurs, cibles, et capteurs. En utilisant comme critère d’optimisation la borne de Cramér-Rao sur la précision de la localisation de cibles multiples, nous proposons une méthode fournissant des solutions approchées aux problèmes d’allocation optimale de puissances seules, de largeurs de bande seules, ou au problème d’allocation conjointe de puissances et de largeurs de bande. Ces solutions sont obtenues en minimisant une suite de problèmes convexes. La qualité de ces solutions approchées est évaluée au travers de nombreuses simulations numériques, mais également par la comparaison avec une borne inférieure définie comme la solution d’un problème d’optimisation avec contraintes relaxées, cette borne pouvant être calculée de façon exacte (numériquement). Cette comparaison permet de constater la proximité de la solution approchée fournie par l’algorithme proposé par rapport à la solution théorique. D’autre part, les simulations ont montré que l’allocation de bande joue un rôle plus important dans les performances de localisation que l’allocation de puissance. Dans la seconde partie de la thèse, on considère le cas de la localisation passive de sources multiples dans un environnement multi-trajet. Ce problème se rencontre notamment dans le cadre de la géolocalisation indoor ou outdoor. Dans ce cas de figure, les approches généralement proposées dans la littérature sont basées sur une méthode ad-hoc de réduction d’interférence couplée à une localisation indirecte obtenue par une estimation de paramètres comme les temps d’arrivée des signaux ou les différences de temps d’arrivée, ou la puissance des signaux reçus. Cependant, les performances de ces approches sont limitées, notamment par le fait que la localisation indirecte d’une cible donnée ne prend pas en compte le fait que les signaux reçus par les différents capteurs émanent d’une seule et même source. Dans cette thèse, nous proposons une modélisation parcimonieuse des signaux reçus. Cette modélisation nous permet, en supposant les formes d’onde connues mais les canaux multi-trajets totalement inconnus, de développer une méthode de localisation directe de l’ensemble des cibles. Cette approche exploite certaines propriétés des canaux, qui permettent de séparer les trajets directs des trajets indirects. Un algorithme d’optimisation conique de second ordre est développé afin d’obtenir une décomposition dite atomique optimale, qui permet d’obtenir une localisation de très bonne précision dans des conditions de propagation difficiles, présentant un phénomène de multi-trajet important et/ou une absence de trajets directs. / Active and passive localization employing widely distributed sensors is a problem of interest in various fields. In active localization, such as in MIMO radar, transmitters emit signals that are reflected by the targets and collected by the receive sensors, whereas, in passive localization the sensors collect the signals emitted by the sources themselves. This dissertation studies optimization methods for high precision active and passive localization. In the case of active localization, multiple transmit elements illuminate the targets from different directions. The signals emitted by the transmitters may differ in power and bandwidth. Such resources are often limited and distributed uniformly among the transmitters. However, previous studies based on the well known Crámer-Rao lower bound have shown that the localization accuracy depends on the locations of the transmitters as well as the individual channel gains between different transmitters, targets and receivers. Thus, it is natural to ask whether localization accuracy may be improved by judiciously allocating such limited resources among the transmitters. Using the Crámer-Rao lower bound for target localization of multiple targets as a figure of merit, approximate solutions are proposed to the problems of optimal power, optimal bandwidth and optimal joint power and bandwidth allocation. These solutions are computed by minimizing a sequence of convex problems. The quality of these solutions is assessed through extensive numerical simulations and with the help of a lower-bound that certifies their optimality. Simulation results reveal that bandwidth allocation policies have a stronger impact on performance than power. Passive localization of radio frequency sources over multipath channels is a difficult problem arising in applications such as outdoor or indoor geolocation. Common approaches that combine ad-hoc methods for multipath mitigation with indirect localization relying on intermediary parameters such as time-of-arrivals, time difference of arrivals or received signal strengths, are unsatisfactory. This dissertation models the localization of known waveforms over unknown multipath channels in a sparse framework, and develops a direct approach in which multiple sources are localized jointly, directly from observations obtained at distributed sources. The proposed approach exploits channel properties that enable to distinguish line-of-sight (LOS) from non-LOS signal paths. Theoretical guarantees are established for correct recovery of the sources’ locations by atomic norm minimization. A second-order-cone-based algorithm is developed to produce the optimal atomic decomposition, and it is shown to produce high accuracy location estimates over complex scenes, in which sources are subject to diverse multipath conditions, including lack of LOS.