La communication mobile est considérée comme l'un des piliers des villes intelligentes, où les citoyens devraient pouvoir bénéficier des services de télécommunications partout et quand ils les souhaitent, d'une manière sûre et peu coûteuse. Cela est possible grâce à un déploiement dense des réseaux mobiles à large bande de dernière génération. Ce déploiement dense entraînera une consommation énergétique plus élevée et donc plus d'émissions de gaz et de pollution. Par conséquent, il est crucial d'un point de vue environnemental de réduire la consommation d'énergie. Dans le cadre de cette thèse, nous introduisons des méthodes dynamiques de gestion de ressources permettant d'augmenter le débit et l'efficacité énergétique, et réduisant ainsi la pollution. Ainsi, nous ciblons les réseaux multicellulaires verts où l'augmentation de l'efficacité énergétique doit tenir en compte de l'accroissement de la demande de débit par les utilisateurs mobiles. Cette augmentation, exponentielle en terme de débit, a poussé les opérateurs à utiliser la totalité du spectre fréquentiel dans toutes les cellules des réseaux mobiles de dernière génération. Par conséquence, l'interférence intercellulaire (ICI : Inter-Cell Interference) devient prépondérante et dégrade la performance des utilisateurs, en particulier ceux ayant de mauvaises conditions radios. Dans cette thèse, nous nous focalisons sur la technique du contrôle de puissance considérée comme une des méthodes clé de la coordination d'interférence Intercellulaire (ICIC : Inter-Cell Interference Coordination), tout en mettant l'accent sur des méthodes efficaces énergétiquement. Nous formulons ce problème d'allocation de la puissance, sur le lien descendant en mettant en œuvre des méthodes centralisées et décentralisées: les méthodes centralisées ayant recours à l'optimisation convexe alors que les méthodes décentralisées se basant sur la théorie des jeux non-coopératifs. Par ailleurs nous proposons ensuite une heuristique de contrôle de puissance qui a l'avantage d'être stable et basée sur des messages de signalisation déjà existant dans le système. Cette heuristique permet d'éviter le gaspillage de la bande passante par des signalisations intercellulaires et de réduire le ICI. De plus, le problème de contrôle de puissance a un impact important sur l'allocation des ressources radios et sur l'association des utilisateurs mobiles à une station de base. Ainsi, dans la deuxième partie de la thèse, nous avons formulé un problème globale englobant le contrôle de puissance, le contrôle d'allocation de ressources radios, et le contrôle de l'association des utilisateurs à une station de base, cela afin d'obtenir une solution globalement efficace. Ces trois sous problèmes sont traités itérativement jusqu'à convergence de la solution globale. En particulier nous proposons pour la problématique d'association des utilisateurs trois algorithmes: un algorithme centralisé, un algorithme semi-distribué et finalement un algorithme complètement distribué se basant sur l'apprentissage par renforcement. Par ailleurs, pour l'allocation de puissance, nous implémentons des solutions centralisées et des solutions distribuées. Les preuves de convergence des algorithmes ont été établies et les simulations approfondies ont permis d'évaluer et de comparer quantitativement les performances, l'efficacité énergétique et le temps de convergence des algorithmes proposés. / Mobile communication is considered as one of the building blocks of smart cities, where citizens should be able to benefit from telecommunications services, wherever they are, whenever they want, and in a secure and non-costly way. This can be done by dense deployment of the latest generation of mobile broadband networks. However, this dense deployment will lead to higher energy consumption, and thus more gas emission and pollution. Therefore, it is crucial from environmental point of view to propose solution reducing energy consumption. In this thesis, we introduce dynamic resource management methods that increase throughput and energy efficiency, and thus reduce pollution. In this framework, we are targeting green multi-cell networks where increased energy efficiency must take into account the increased demand of data by mobile users. This increase, which is exponential in terms of throughput, pushed operators to use the entire frequency spectrum in all cells of the latest generation of mobile networks. As a result, Inter-Cellular Interference (ICI) became preponderant and degraded the performance of users, particularly those with poor radio conditions. In this thesis, we focus on the techniques of power control on the downlink direction, which is considered as one of the key methods of Inter-Cell Interference Coordination (ICIC) while focusing on energy efficient methods. We propose centralized and decentralized methods for this problem of power allocation: centralized methods through convex optimization, and decentralized methods based on non-cooperative game theory. Furthermore, we propose a power control heuristic which has the advantage of being stable and based on signaling messages already existing in the system. The power control problem has a relevant impact on the allocation of radio resources and on the association of mobile users with their servicing Base Station. Therefore, in the second part of the thesis, we formulated a global problem encompassing power control, radio resources allocation, and control of users’ association to a base station. These three sub-problems are treated iteratively until the convergence to the overall solution. In particular, we propose three algorithms for the user association problem: a centralized algorithm, a semi-distributed algorithm and finally a fully distributed algorithm based on reinforcement learning. In addition, for power allocation we implement centralized solutions and distributed solutions. The proof of convergence for the various algorithms is established and the in-depth simulations allow us to evaluate and compare quantitatively the performance, the energy efficiency, and the convergence time of the proposed algorithms.

No description available.

No description available.

No description available.

Control Allocation addresses the problem of the management of multiple, redundant control effectors. Generally speaking, control allocation is any method that is used to determine how the controls of a system should be positioned to achieve some desired effect. An infinite number of allocation methods exist, from the straight-forward direct allocation technique, to the daisy chaining approach, to the computationally simple generalized inverse method. Because different methods have advantages and disadvantages with respect to others, the determination of the "optimal" control allocation method is left to the system designer. The many tradeoffs that are addressed during control system design, of which control allocation is an integral part, dictate the need for a reliable, computer-based design tool. The Control Allocation Toolbox for MATLAB satisfies such a need by providing the designer with a means of testing/comparing the validity of certain allocation methods under prescribed conditions. The issues involved in the development and implementation of the Control Allocation Toolbox are discussed. / Master of Science

The manning of a ship is a major driver of life cycle cost. The U.S. Government Accounting Office (GAO) has determined that manpower is the single most influential component in the life cycle cost of a ship. Life cycle cost is largely determined by decisions made during concept design. Consequently, reliable manpower estimates need to be included early in the design process, preferably in concept design. The ship concept exploration process developed at Virginia Tech uses a Multi-Objective Genetic Optimization to search the design space for feasible and non-dominated ship concepts based on cost, risk and effectiveness. This requires assessment of thousands of designs without human intervention. The total ship design problem must be set up before actually running the optimization. If manning is to be included in this process, manning estimate tools must be run seamlessly as part of the overall ship synthesis and optimization. This thesis provides a method of implementing a manning task network analysis tool (ISMAT, Integrated Simulation Manning Analysis Tool, Micro Analysis and Design) in an overall ship synthesis program and design optimization. The inputs to the analysis are ship systems (propulsion, combat systems, communication, etc), maintenance strategy, and level of automation. The output of the manning model is the number of crew required to accomplish a given mission for a particular selection of systems, maintenance and automation. Task network analysis programs are ideal for this problem. They can manage the probabilistic nature of a military mission and equipment maintenance, and can be used to simplify the problem by breaking down the complex functions and tasks of a ship's crew. The program builds large and complex functions from small related tasks. This simplifies the calculation of personnel and time utilization, and allows a more flexible scheme for building complex mission scenarios. In this thesis, ISMAT is run in a pre-optimization step to build a response surface model (RSM) for calculating required manning as a function of systems, maintenance and automation. The RSM is added to the ship synthesis model to calculate required manning, and a concept exploration case study is performed for an Air Superiority Cruiser (CGX) using this model. The performance of the manning model in this case study is assessed and recommendations are made for future work. This research shows that there is a difference between minimum manning and optimal manning on US Navy Ships. / Master of Science

We study a wireless full-duplex network with imperfect interference cancellation and solve the routing and power allocation problem in this network. We use a model that focuses on the effects of full-duplex by including residual self-interference and one hop interference while other interfering signals are considered negligible in comparison. We first solve the optimal power allocation for a fixed route. We then propose a priority-first search algorithm to find the joint route and power allocation to maximize throughput. The algorithm proposed has a non decomposable priority metric, but is efficiently evaluated by our solution for a fixed route. We analyze the performance of our solution in a more realistic model by deriving bounds between optimal solutions in both models. Through simulations we show that, even with imperfect interference cancellation, full-duplex achieves a higher throughput than half-duplex or direct transmission for moderate transmission power.

In this thesis a distributed, dynamic, load aware, joint power and frequency allocation protocol for 4G networks along with system-level simulated results are presented. In all cellular networks, a key limiting factor for throughput is inter-cell interference, especially at the cell edges. Several methods have been proposed and adopted in each mobile network generation to cancel or suppress its effects, whereas each method has its drawbacks in terms of receiver complexity or additional control nodes. However, the proposed protocol presented here does not impose any architectural changes.  In 4G networks such as LTE, the choice of OFDMA for the air interface has paved the way for selective frequency and power allocation in the available spectrum. Taking advantage of this opportunity, fractional frequency reuse (FFR) has been proposed in OFDMA based mobile networks in order to reduce the throughput loss at the cell edges due to inter-cell interference. In FFR, center users lose part of available spectrum that is dedicated to the edge users.  Our protocol aims to minimize this loss of center users incurred by FFR, at the cost of minimal degradation at the edges. An eNodeB, only when overloaded, requests its neighbours’ edge band to be used for its center users at a reduced power level. This is done via small message exchange between the eNodeBs. The neighbors of the overloaded eNodeBs solve a small local knapsack problem to decide whether band lending is feasible or not. A distinguishing feature of this protocol is the power level adjustment for the borrowed band, which is mutually decided by the borrower and lender. The band is released when it is not needed or it is causing unacceptable loss to the lender. The implementation is done in a Matlab based LTE system level simulator. For the implementation of our protocol in the simulator, starting point was implementation of FFR-3 functionality, a prerequisite and a baseline for comparison with our protocol. Results are compared among three different setups of re-use1, FFR-3 and our protocol by varying number of overloaded eNodeBs for various numbers of scenarios and the comparison is made based on the center users’ throughput, edge users’ throughput. An estimation of time and protocol overhead is also presented. We have observed center users’ throughput gain up to 46%, at the cost of 9% edge users’ throughput loss, when compared to the classic FFR-3 scheme. The overall system throughput goes up to 26 % in heavily loaded scenario.   The utility of the protocol for an LTE system is evident from the results, which is supported by the dynamic and decentralized nature of the protocol. This ensures better utilization of spectrum, by temporarily allocating more bandwidth where it is needed more.

This study is based on the strategic and logistical challenges of having a complex distribution network, which can make it difficult to get a holistic view over the distribution costs. The costs are often aggregated for many products, which makes it challenging to use as decision support on a product level. Many companies lack a tool to handle this complexity, since the costs and profitability varies between the channels and intermediaries used. This makes it problematic to determine the profitability on a product level. In the different parts of the distribution chain, there are elements that drive the costs for each activity, called activity drivers. When these activity drivers have been identified, they can be used to allocate the distribution costs to the different products. The aim of this study is to develop a tool that can be used to categorize distribution costs and to determine which activity drivers that result in the fairest cost allocation. The fairest cost allocation is a complex expression, and is briefly defined as the allocation key that result in a costs allocation that represent each products level of resource consumption. This means that products that have consumed a large amount of resources should carry a larger part of the costs compared to the products that have consumed a smaller amount of resources. Sometimes it is not obvious which allocation key that represents the reality in the fairest way, and in that case, the allocated costs are compared to the products sales values. The sales value often differs between the products. The determined allocation key is the one that result in the most even allocation when comparing the allocated cost to the sales values. The case company Swedish Orphan Biovitrum (Sobi) is located in Stockholm, Sweden. They find it difficult to get a view over the costs for the different parts of the distribution chain, and to allocate the costs fair between the products. This study have investigated the distribution from Sobi’s central warehouse in the Netherlands to the end customers in France, Germany, Italy, Spain, the United Kingdom via the local storages in each country, as well as Sweden and Estonia. This was done by categorizing the costs for each activity in the invoices from the local storages, into different cost categories. After this, the costs were allocated with different allocation keys that thereafter were compared, to find the most fair allocation key per category. In the end of this study, the lessons learned and methods used have been written down, and an allocation tool has been developed. Any company that wants to make strategic decisions on a product level can use this tool. Throughout the study, the five steps that make up the allocation tool have been followed. The tool is divided into the following steps; determination of cost categories, choice of activities, selection of activity drivers, categorization of costs and analyzing activity drivers. When choosing allocation key, it is essential to find the balance between an even allocation of the costs between the products, and to make sure that the allocation represent each products level of resource consumption. If the allocation is unfair, it can make products look unprofitable, even though they actually are profitable and necessary in reality. The difficulties to find a balance show the complexity in the determination of the most fair allocation key, since it is not always obvious. If the cost categories had been divided into smaller categories with more similar activity drivers, the dilemma of choosing allocation key might have been solved. However, it is important to bear in mind that when using more cost categories, the categorization and allocation becomes more time consuming. The tool has been created as a result of this study, and is based on a complex situation, which means that assumptions and simplifications have been made to be able to draw general conclusions. It is important to bare these simplifications in mind, when applying the tool to other situations than the one investigated in this study. The allocation tool can be used to draw strategic conclusions on a product level, since it makes it possible to be aware of the profitability of the products and, if necessary, exclude unprofitable products from the product assortment.

No description available.