Les travaux de cette thèse s’articulent autour des « problèmes d’ordonnancement multiagent avec une fonction objectif globale ». Ces modèles considèrent différents agents associés à des sous-ensembles de travaux disjoints, chacun d’eux vise à minimiser un objectif qui ne dépend que de ses propres travaux. Un critère global est aussi considéré, qui est appliqué à la totalité des travaux. La résolution de ces problèmes revient à trouver les meilleurs compromis entre les critères des agents et le critère global. Ces problèmes sont une classe particulière des problèmes d’ordonnancement « multi-agents » qui ont connu une grande expansion, reflétant leurs intérêts dans le domaine de l’ordonnancement. / This thesis addresses the multi-agent scheduling problems with a global objective function. We consider the problems featured by various agents, each of which is associated with a distinct subset of jobs. Each agent aims at minimizing a certain objective function, which only operates on its assigned jobs. A global criterion associated with a global agent is applied on the whole set of the jobs. Solving these problems involves finding the best compromises between the requirements of agents and that of the global agent. These problems belong to a particular class of multi-criteria scheduling problems. Such a class has drawn a significant interest to researchers in the area of scheduling and operational research.

A Virtualização de Funções de Rede (NFV – Network Function Virtualization) é um novo conceito arquitetural que está remodelando a operação de funções de rede (e.g., firewall, gateways e proxies). O conceito principal de NFV consiste em desacoplar a lógica de funções de rede dos dispositivos de hardware especializados e, desta forma, permite a execução de imagens de software sobre hardware de prateleira (COTS – Commercial Off-The-Shelf). NFV tem o potencial para tornar a operação das funções de rede mais flexíveis e econômicas, primordiais em ambientes onde o número de funções implantadas pode chegar facilmente à ordem de centenas. Apesar da intensa atividade de pesquisa na área, o problema de posicionar e encadear funções de rede virtuais (VNF – Virtual Network Functions) de maneira escalável e com baixo custo ainda apresenta uma série de limitações. Mais especificamente, as estratégias existentes na literatura negligenciam o aspecto de encadeamento de VNFs (i.e., objetivam sobretudo o posicionamento), não escalam para o tamanho das infraestruturas NFV (i.e., milhares de nós com capacidade de computação) e, por último, baseiam a qualidade das soluções obtidas em custos operacionais não representativos. Nesta tese, aborda-se o posicionamento e o encadeamento de funções de rede virtualizadas (VNFPC – Virtual Network Function Placement and Chaining) como um problema de otimização no contexto intra- e inter-datacenter. Primeiro, formaliza-se o problema VNFPC e propõe-se um modelo de Programação Linear Inteira (ILP) para resolvêlo. O objetivo consiste em minimizar a alocação de recursos, ao mesmo tempo que atende aos requisitos e restrições de fluxo de rede. Segundo, aborda-se a escalabilidade do problema VNFPC para resolver grandes instâncias do problema (i.e., milhares de nós NFV). Propõe-se um um algoritmo heurístico baseado em fix-and-optimize que incorpora a meta-heurística Variable Neighborhood Search (VNS) para explorar eficientemente o espaço de solução do problema VNFPC. Terceiro, avalia-se as limitações de desempenho e os custos operacionais de estratégias típicas de aprovisionamento ambientes reais de NFV. Com base nos resultados empíricos coletados, propõe-se um modelo analítico que estima com alta precisão os custos operacionais para requisitos de VNFs arbitrários. Quarto, desenvolve-se um mecanismo para a implantação de encadeamentos de VNFs no contexto intra-datacenter. O algoritmo proposto (OCM – Operational Cost Minimization) baseia-se em uma extensão da redução bem conhecida do problema de emparelhamento ponderado (i.e., weighted perfect matching problem) para o problema de fluxo de custo mínimo (i.e., min-cost flow problem) e considera o desempenho das VNFs (e.g., requisitos de CPU), bem como os custos operacionais estimados. Os resultados alcaçados mostram que o modelo ILP proposto para o problema VNFPC reduz em até 25% nos atrasos fim-a-fim (em comparação com os encadeamentos observados nas infra-estruturas tradicionais) com um excesso de provisionamento de recursos aceitável – limitado a 4%. Além disso, os resultados evidenciam que a heurística proposta (baseada em fix-and-optimize) é capaz de encontrar soluções factíveis de alta qualidade de forma eficiente, mesmo em cenários com milhares de VNFs. Além disso, provê-se um melhor entendimento sobre as métricas de desempenho de rede (e.g., vazão, consumo de CPU e capacidade de processamento de pacotes) para as estratégias típicas de implantação de VNFs adotadas infraestruturas NFV. Por último, o algoritmo proposto no contexto intra-datacenter (i.e. OCM) reduz significativamente os custos operacionais quando comparado aos mecanismos de posicionamento típicos uti / Network Function Virtualization (NFV) is a novel concept that is reshaping the middlebox arena, shifting network functions (e.g. firewall, gateways, proxies) from specialized hardware appliances to software images running on commodity hardware. This concept has potential to make network function provision and operation more flexible and cost-effective, paramount in a world where deployed middleboxes may easily reach the order of hundreds. Despite recent research activity in the field, little has been done towards scalable and cost-efficient placement & chaining of virtual network functions (VNFs) – a key feature for the effective success of NFV. More specifically, existing strategies have neglected the chaining aspect of NFV (focusing on efficient placement only), failed to scale to hundreds of network functions and relied on unrealistic operational costs. In this thesis, we approach VNF placement and chaining as an optimization problem in the context of Inter- and Intra-datacenter. First, we formalize the Virtual Network Function Placement and Chaining (VNFPC) problem and propose an Integer Linear Programming (ILP) model to solve it. The goal is to minimize required resource allocation, while meeting network flow requirements and constraints. Then, we address scalability of VNFPC problem to solve large instances (i.e., thousands of NFV nodes) by proposing a fixand- optimize-based heuristic algorithm for tackling it. Our algorithm incorporates a Variable Neighborhood Search (VNS) meta-heuristic, for efficiently exploring the placement and chaining solution space. Further, we assess the performance limitations of typical NFV-based deployments and the incurred operational costs of commodity servers and propose an analytical model that accurately predict the operational costs for arbitrary service chain requirements. Then, we develop a general service chain intra-datacenter deployment mechanism (named OCM – Operational Cost Minimization) that considers both the actual performance of the service chains (e.g., CPU requirements) as well as the operational incurred cost. Our novel algorithm is based on an extension of the well-known reduction from weighted matching to min-cost flow problem. Finally, we tackle the problem of monitoring service chains in NFV-based environments. For that, we introduce the DNM (Distributed Network Monitoring) problem and propose an optimization model to solve it. DNM allows service chain segments to be independently monitored, which allows specialized network monitoring requirements to be met in a efficient and coordinated way. Results show that the proposed ILP model for the VNFPC problem leads to a reduction of up to 25% in end-to-end delays (in comparison to chainings observed in traditional infrastructures) and an acceptable resource over-provisioning limited to 4%. Also, we provide strong evidences that our fix-and-optimize based heuristic is able to find feasible, high-quality solutions efficiently, even in scenarios scaling to thousands of VNFs. Further, we provide indepth insights on network performance metrics (such as throughput, CPU utilization and packet processing) and its current limitations while considering typical deployment strategies. Our OCM algorithm reduces significantly operational costs when compared to the de-facto standard placement mechanisms used in Cloud systems. Last, our DNM model allows finer grained network monitoring with limited overheads. By coordinating the placement of monitoring sinks and the forwarding of network monitoring traffic, DNM can reduce the number of monitoring sinks and the network resource consumption (54% lower than a traditional method).

Orientador: Marcos Júlio Rider Flores / Resumo: Nesta tese são apresentados três modelos de programação matemática que abordam os problemas de otimização relacionados ao gerenciamento da energia nos sistemas de distribuição de energia elétrica (SDEE), como: 1) Programação ótima das entregas e carregamento dos veículos elétricos (VEs) durante a navegação em um mapa de cidade, 2) Gerenciamento ótimo pelo lado da demanda considerando um sistema fotovoltaico híbrido (SFH) em uma residência em baixa tensão (RBT) no SDEE, e 3) O melhoramento do fator de carga (FC) do SDEE através do controle da demanda. O primeiro problema visa minimizar os custos relacionados com a manutenção e geração de horas extra durante a operação de uma frota de VEs, levando em conta um conjunto de entregas pre-especificadas, assim como, pontos de carregamento alocados ao longo de cada via urbana (principal e/ou secundária) pertencente ao mapa da cidade. No segundo problema, para uma residência em baixa tensão é planejado um perfil ótimo de consumo para o dia seguinte. Este perfil de consumo é obtido através de um programa de gerenciamento pelo lado da demanda (GLD) que considera uma estrutura tarifária e um esquema de operação que otimiza os recursos energéticos vindos de um SFH e o SDEE. Para cada problema de otimização é apresentado o seu correspondente modelo de programação não linear inteiro misto (PNLIM). O terceiro problema visa minimizar os custos por compra de energia (consumo e perdas de potência ativa) da concessionária, levando em conta, o co... (Resumo completo, clicar acesso eletrônico abaixo) / Doutor

Submitted by Caroline Periotto (carol@ufscar.br) on 2016-09-19T18:29:26Z No. of bitstreams: 1 DissRBL.pdf: 2233937 bytes, checksum: 7a73fb1e4ba8f340cb8d4bdf4e8153d4 (MD5) / Approved for entry into archive by Marina Freitas (marinapf@ufscar.br) on 2016-09-20T18:01:34Z (GMT) No. of bitstreams: 1 DissRBL.pdf: 2233937 bytes, checksum: 7a73fb1e4ba8f340cb8d4bdf4e8153d4 (MD5) / Approved for entry into archive by Marina Freitas (marinapf@ufscar.br) on 2016-09-20T18:05:15Z (GMT) No. of bitstreams: 1 DissRBL.pdf: 2233937 bytes, checksum: 7a73fb1e4ba8f340cb8d4bdf4e8153d4 (MD5) / Made available in DSpace on 2016-09-20T18:05:28Z (GMT). No. of bitstreams: 1 DissRBL.pdf: 2233937 bytes, checksum: 7a73fb1e4ba8f340cb8d4bdf4e8153d4 (MD5) Previous issue date: 2015-12-14 / Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) / The Brazilian market for dairy products is highly competitive and is still dominated by multinationals. Thus, it's essential that the national companies to invest in operational efficiency in order to compete successfully. Therefore, this research focuses in a small dairy company of the state of Pará, whose distribution system requires periodic stops for battery recharging. Moreover, the system is characterized by multiple periods and multiple time windows. To our knowledge, so far the literature hasn't presented a directly applicable methodology for the treatmeant of the application with similar characteristics. Therefore, aiming to provide more effective solutions than the ones in practice, a mixed integer linear model was developed to describe (and solve) the problem as a vehicle routing problem with time window constraints, multi-period and periodic stops for recharging. The results showed that the model adequately describes the distribution of the company, and the obtained solutions are better than those currently practiced. Furthermore, the model shows good performance within 3600 seconds of computational time for instances of 40 customers, 1 vehicle and 1 and 2 recharging stations. Aiming to tackle with larger examples, we developed a mathematical programming heuristic Relax-and-Fix to solve the model. It was also tested adapting a similar model present in the literature in order to analyze if the relaxation of some restrictions have a positive impact on the quality of the solutions. Finally, we propose a location-distribution model for recharging stations in order to examine whether the addition of new stations produce significant improvements in the solutions. / O mercado brasileiro de laticínios é altamente competitivo, sendo ainda dominado por multinacionais. Desta forma, é indispensável que as empresas nacionais invistam na eficiência operacional de forma a concorrer satisfatoriamente neste mercado. À vista disso, o foco desta pesquisa é uma pequena empresa de laticínios do Estado do Pará, cujo o sistema de distribuição a pontos de venda possui a particularidade de necessitar de paradas periódicas para recarga da bateria. Ainda, esse sistema tem como característica múltiplos períodos e múltiplas janelas de tempo. Nota-se que o levantamento bibliográfico realizado não apontou nenhuma metodologia diretamente aplicável para solução da aplicação real tratada. Portanto, objetivando provar soluções mais efetivas que as correntemente aplicadas na prática, foi desenvolvido um modelo linear inteiro misto para descrever (e resolver) o problema como um problema de roteamento de veículos com restrições janela de tempo, multi-períodos e paradas periódicas para recarga. Os resultados mostraram que o modelo proposto descreve adequadamente o processo distribuição da empresa, e as soluções obtidas são melhores que as praticadas atualmente. Além disso, verificou-se que o modelo tem um desempenho satisfatório em até 3600 segundos de processamento para instâncias de até 40 clientes, 1 veículo e até 2 postos de recarga. Com vistas ao tratamento de problemas de maior porte, foi desenvolvida uma heurística de programação matemática do tipo Relax-and-Fix para resolução do modelo. Também foi testada a adaptação de um modelo similar presente na literatura, visando analisar se a relaxação de algumas restrições impacta positivamente na qualidade das soluções. Finalmente, é proposto um modelo de localização-distribuição de postos de recarga, de forma a se examinar se a inclusão de novos postos produz melhorias significativas nas soluções atuais. / FAPESP: 14/10330-0

Submitted by Aelson Maciera (aelsoncm@terra.com.br) on 2017-05-22T20:04:01Z No. of bitstreams: 1 TeseALS.pdf: 2588373 bytes, checksum: 85ce391accc868c711a33c83006ee738 (MD5) / Approved for entry into archive by Ronildo Prado (ronisp@ufscar.br) on 2017-05-31T13:43:27Z (GMT) No. of bitstreams: 1 TeseALS.pdf: 2588373 bytes, checksum: 85ce391accc868c711a33c83006ee738 (MD5) / Approved for entry into archive by Ronildo Prado (ronisp@ufscar.br) on 2017-05-31T13:43:34Z (GMT) No. of bitstreams: 1 TeseALS.pdf: 2588373 bytes, checksum: 85ce391accc868c711a33c83006ee738 (MD5) / Made available in DSpace on 2017-05-31T13:52:08Z (GMT). No. of bitstreams: 1 TeseALS.pdf: 2588373 bytes, checksum: 85ce391accc868c711a33c83006ee738 (MD5) Previous issue date: 2017-03-07 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) / Vehicle routing problems occur in many practical situations where the distribution of goods and / or services to different demand points is necessary. In this context, this research aims to study a ship routing and scheduling problem that arises at the collection and delivery operations of different types of crude oil from offshore platforms to coastal terminals. In the paradigm adopted for the representation of the problem, the transportation is largely the result of the need to maintain inventories at each supply point (platforms) between minimum and maximum levels, considering production rates on these operating points and the demand attendance of each product in the coastal terminals. The routing and scheduling of the fleet aims to achieve minimum variable cost solutions, and considers various operational constraints, such as the maximum cargo volume transported on each ship, the ships mooring in the operational points ports, the simultaneous unloading of the ships in terminals with more than one berth, among many others. In this research, Inventory Constrained Routing Problem (ICRP) models in the maritime context have been modified and extended for appropriately representating and solving real problems based on data collected in a case study performed on a Brazilian oil company, involving relatively short distances and time horizons. Small sized instances are solved by a mathematical programming software. Given the difficulties of solving larger examples, this study proposes a multistart heuristic method that includes a metaheuristic GRASP and improvement procedures, and also a rolling horizon heuristic. Both methods provide feasible good quality solutions in reasonable computing times. In order to improve the quality of the solutions found by these constructive methods, it is also discussed a procedure that combines the mathematical programming software and local search heuristic methods (matheuristic). The results show the potential of the proposed models and solution methods to tackle the problem and produce competitive solutions. / Problemas de roteirização de veículos ocorrem em diversas situações práticas onde se faz necessária a distribuição de bens e/ou serviços a pontos dispersos de demanda. Nesse contexto, a presente pesquisa visa o estudo de um problema de roteirização e programação de navios presente em operações de coleta e entrega de diferentes tipos de óleo cru de diversas plataformas offshore para vários terminais costeiros. No paradigma adotado para representação do problema, o transporte dos produtos é em grande parte o resultado da necessidade de manutenção dos estoques em cada ponto de suprimento (plataformas) entre níveis mínimos e máximos, considerando-se as taxas de produção nesses pontos operacionais, assim como o atendimento da demanda de cada produto nos terminais costeiros para abastecer as refinarias. A roteirização e programação da frota visa a obtenção de soluções de mínimo custo variável e considera várias restrições operacionais, tais como o volume máximo de carga transportada em cada navio, a viabilidade de atracação de navios em portos dos pontos operacionais, os descarregamentos simultâneos de navios em terminais com mais de um berço, dentre várias outras. Nesse sentido, modelos de otimização da literatura de roteirização veículos com restrições de estoque (Inventory Constrained Routing Problem – ICRP) no contexto marítimo foram modificados e estendidos para representação do problema e resolução de exemplares de uma situação real, definidos a partir de dados coletados em um estudo de caso realizado em uma empresa petrolífera nacional, envolvendo distâncias relativamente curtas e com horizontes de planejamento de curto prazo (poucas semanas). Exemplares de pequeno porte são resolvidos por meio da utilização de um software de programação matemática. Dada a dificuldade de resolução dos exemplos de maior porte, é proposto um método heurístico de múltiplos reinícios composto por uma metaheurística GRASP e procedimentos de melhoria, além de uma heurística de horizonte rolante, que proporcionaram a obtenção de soluções factíveis de boa qualidade em tempos computacionais aceitáveis. Com intuito de melhorar a qualidade das soluções encontradas pelos métodos construtivos, é também discutido um procedimento que combina o software de programação matemática e métodos heurísticos com busca local (mateheurística). Os resultados mostram o potencial dos modelos e métodos de solução aqui desenvolvidos e propostos para abordar o problema e produzir soluções competitivas em relação às soluções da empresa.

The struggle to model and solve Combinatorial Optimization Problems (COPs) has challenged the development of new approaches to deal with COPs. In one of the front lines of such approaches, Operational Research (OR) and Constraint Programming (CP) optimization techniques are beginning to converge, despite their very different origins. More specifically, Mixed Integer Linear Programming (MILP) and Constraint Logic Programming (CLP) are at the confluence of the OR and the CP fields. This thesis summarizes and contrasts the essential characteristics of MILP and CLP, and the ways that they can be fruitfully combined. Chapters 1 to 3 sketch the intellectual background for recent efforts at integration and the main results achieved. In addition, these chapters highlight that CLP is known by its reach modeling framework, and the MILP modeling vocabulary is just based on inequalities, which makes the modeling process hard and error-prone. Therefore, a combined CLP-MILP approach suffers from this MILP inherited drawback. In chapter 4, this issue is addressed, and some "high-level" MILP modeling structures based on logical inference paradigms are proposed. These structures help the formulation of MILP models, and can be seen as a contribution towards a unifying modeling framework for a combined CLP-MILP approach. In addition, chapter 5 presents an MILP formulation addressing a combinatorial problem. This problem is focused on issues regarding the oil industry, more specifically, issues involving the scheduling of operational activities in a multi-product pipeline. Chapter 5 demonstrates the applicability of the high-level MILP modeling structures in a real-world scenario. Furthermore, chapter 6 presents a CLP-MILP formulation addressing the same scheduling problem previously exploited. This chapter demonstrates the applicability of the high-level MILP modeling structures in an integrated CLP-MILP modeling framework. The set of simulations conducted indicates that the combined CLP-MILP model was solved to optimality faster than either the MILP model or the CLP model. Thus, the CLP-MILP framework is a promising alternative to deal with the computational burden of this pipeline-scheduling problem. In essence, this thesis considers the integration of CLP and MILP in a modeling standpoint: it conveys the fundamentals of both techniques and the modeling features that help establish a combined CLP-MILP approach. Herein, the concentration is on the building of MILP and CLP-MILP models rather than on the solution process.

The struggle to model and solve Combinatorial Optimization Problems (COPs) has challenged the development of new approaches to deal with COPs. In one of the front lines of such approaches, Operational Research (OR) and Constraint Programming (CP) optimization techniques are beginning to converge, despite their very different origins. More specifically, Mixed Integer Linear Programming (MILP) and Constraint Logic Programming (CLP) are at the confluence of the OR and the CP fields. This thesis summarizes and contrasts the essential characteristics of MILP and CLP, and the ways that they can be fruitfully combined. Chapters 1 to 3 sketch the intellectual background for recent efforts at integration and the main results achieved. In addition, these chapters highlight that CLP is known by its reach modeling framework, and the MILP modeling vocabulary is just based on inequalities, which makes the modeling process hard and error-prone. Therefore, a combined CLP-MILP approach suffers from this MILP inherited drawback. In chapter 4, this issue is addressed, and some "high-level" MILP modeling structures based on logical inference paradigms are proposed. These structures help the formulation of MILP models, and can be seen as a contribution towards a unifying modeling framework for a combined CLP-MILP approach. In addition, chapter 5 presents an MILP formulation addressing a combinatorial problem. This problem is focused on issues regarding the oil industry, more specifically, issues involving the scheduling of operational activities in a multi-product pipeline. Chapter 5 demonstrates the applicability of the high-level MILP modeling structures in a real-world scenario. Furthermore, chapter 6 presents a CLP-MILP formulation addressing the same scheduling problem previously exploited. This chapter demonstrates the applicability of the high-level MILP modeling structures in an integrated CLP-MILP modeling framework. The set of simulations conducted indicates that the combined CLP-MILP model was solved to optimality faster than either the MILP model or the CLP model. Thus, the CLP-MILP framework is a promising alternative to deal with the computational burden of this pipeline-scheduling problem. In essence, this thesis considers the integration of CLP and MILP in a modeling standpoint: it conveys the fundamentals of both techniques and the modeling features that help establish a combined CLP-MILP approach. Herein, the concentration is on the building of MILP and CLP-MILP models rather than on the solution process.

A Virtualização de Funções de Rede (NFV – Network Function Virtualization) é um novo conceito arquitetural que está remodelando a operação de funções de rede (e.g., firewall, gateways e proxies). O conceito principal de NFV consiste em desacoplar a lógica de funções de rede dos dispositivos de hardware especializados e, desta forma, permite a execução de imagens de software sobre hardware de prateleira (COTS – Commercial Off-The-Shelf). NFV tem o potencial para tornar a operação das funções de rede mais flexíveis e econômicas, primordiais em ambientes onde o número de funções implantadas pode chegar facilmente à ordem de centenas. Apesar da intensa atividade de pesquisa na área, o problema de posicionar e encadear funções de rede virtuais (VNF – Virtual Network Functions) de maneira escalável e com baixo custo ainda apresenta uma série de limitações. Mais especificamente, as estratégias existentes na literatura negligenciam o aspecto de encadeamento de VNFs (i.e., objetivam sobretudo o posicionamento), não escalam para o tamanho das infraestruturas NFV (i.e., milhares de nós com capacidade de computação) e, por último, baseiam a qualidade das soluções obtidas em custos operacionais não representativos. Nesta tese, aborda-se o posicionamento e o encadeamento de funções de rede virtualizadas (VNFPC – Virtual Network Function Placement and Chaining) como um problema de otimização no contexto intra- e inter-datacenter. Primeiro, formaliza-se o problema VNFPC e propõe-se um modelo de Programação Linear Inteira (ILP) para resolvêlo. O objetivo consiste em minimizar a alocação de recursos, ao mesmo tempo que atende aos requisitos e restrições de fluxo de rede. Segundo, aborda-se a escalabilidade do problema VNFPC para resolver grandes instâncias do problema (i.e., milhares de nós NFV). Propõe-se um um algoritmo heurístico baseado em fix-and-optimize que incorpora a meta-heurística Variable Neighborhood Search (VNS) para explorar eficientemente o espaço de solução do problema VNFPC. Terceiro, avalia-se as limitações de desempenho e os custos operacionais de estratégias típicas de aprovisionamento ambientes reais de NFV. Com base nos resultados empíricos coletados, propõe-se um modelo analítico que estima com alta precisão os custos operacionais para requisitos de VNFs arbitrários. Quarto, desenvolve-se um mecanismo para a implantação de encadeamentos de VNFs no contexto intra-datacenter. O algoritmo proposto (OCM – Operational Cost Minimization) baseia-se em uma extensão da redução bem conhecida do problema de emparelhamento ponderado (i.e., weighted perfect matching problem) para o problema de fluxo de custo mínimo (i.e., min-cost flow problem) e considera o desempenho das VNFs (e.g., requisitos de CPU), bem como os custos operacionais estimados. Os resultados alcaçados mostram que o modelo ILP proposto para o problema VNFPC reduz em até 25% nos atrasos fim-a-fim (em comparação com os encadeamentos observados nas infra-estruturas tradicionais) com um excesso de provisionamento de recursos aceitável – limitado a 4%. Além disso, os resultados evidenciam que a heurística proposta (baseada em fix-and-optimize) é capaz de encontrar soluções factíveis de alta qualidade de forma eficiente, mesmo em cenários com milhares de VNFs. Além disso, provê-se um melhor entendimento sobre as métricas de desempenho de rede (e.g., vazão, consumo de CPU e capacidade de processamento de pacotes) para as estratégias típicas de implantação de VNFs adotadas infraestruturas NFV. Por último, o algoritmo proposto no contexto intra-datacenter (i.e. OCM) reduz significativamente os custos operacionais quando comparado aos mecanismos de posicionamento típicos uti / Network Function Virtualization (NFV) is a novel concept that is reshaping the middlebox arena, shifting network functions (e.g. firewall, gateways, proxies) from specialized hardware appliances to software images running on commodity hardware. This concept has potential to make network function provision and operation more flexible and cost-effective, paramount in a world where deployed middleboxes may easily reach the order of hundreds. Despite recent research activity in the field, little has been done towards scalable and cost-efficient placement & chaining of virtual network functions (VNFs) – a key feature for the effective success of NFV. More specifically, existing strategies have neglected the chaining aspect of NFV (focusing on efficient placement only), failed to scale to hundreds of network functions and relied on unrealistic operational costs. In this thesis, we approach VNF placement and chaining as an optimization problem in the context of Inter- and Intra-datacenter. First, we formalize the Virtual Network Function Placement and Chaining (VNFPC) problem and propose an Integer Linear Programming (ILP) model to solve it. The goal is to minimize required resource allocation, while meeting network flow requirements and constraints. Then, we address scalability of VNFPC problem to solve large instances (i.e., thousands of NFV nodes) by proposing a fixand- optimize-based heuristic algorithm for tackling it. Our algorithm incorporates a Variable Neighborhood Search (VNS) meta-heuristic, for efficiently exploring the placement and chaining solution space. Further, we assess the performance limitations of typical NFV-based deployments and the incurred operational costs of commodity servers and propose an analytical model that accurately predict the operational costs for arbitrary service chain requirements. Then, we develop a general service chain intra-datacenter deployment mechanism (named OCM – Operational Cost Minimization) that considers both the actual performance of the service chains (e.g., CPU requirements) as well as the operational incurred cost. Our novel algorithm is based on an extension of the well-known reduction from weighted matching to min-cost flow problem. Finally, we tackle the problem of monitoring service chains in NFV-based environments. For that, we introduce the DNM (Distributed Network Monitoring) problem and propose an optimization model to solve it. DNM allows service chain segments to be independently monitored, which allows specialized network monitoring requirements to be met in a efficient and coordinated way. Results show that the proposed ILP model for the VNFPC problem leads to a reduction of up to 25% in end-to-end delays (in comparison to chainings observed in traditional infrastructures) and an acceptable resource over-provisioning limited to 4%. Also, we provide strong evidences that our fix-and-optimize based heuristic is able to find feasible, high-quality solutions efficiently, even in scenarios scaling to thousands of VNFs. Further, we provide indepth insights on network performance metrics (such as throughput, CPU utilization and packet processing) and its current limitations while considering typical deployment strategies. Our OCM algorithm reduces significantly operational costs when compared to the de-facto standard placement mechanisms used in Cloud systems. Last, our DNM model allows finer grained network monitoring with limited overheads. By coordinating the placement of monitoring sinks and the forwarding of network monitoring traffic, DNM can reduce the number of monitoring sinks and the network resource consumption (54% lower than a traditional method).

The thesis deals with recent problems of waste management in the Czech Republic. In connection with the existing software implementation, the author focuses on the gradual development of advanced mathematical programming models, which generalize existing approaches. The author applies acquired knowledge in the areas of network flows, linear, integer, and stochastic programming. The important role is played by modifications and transformations of the discussed models. They are further used to obtain the experimental results for real-world input data by implementation in GAMS.

Designing structures for lightness is an intelligent and responsible way for engineers and architects to conceive structural systems. Lightweight structures are able to bridge wide spans with a least amount of material. However, the quest for lightness remains an utopia without the driving constraints that give sense to contemporary structural design.<p><p>Previously proposed computational methods for designing lightweight structures focused either on finding an equilibrium shape, or are restricted to fairly small design applications. In this work, we aim to develop a general, robust, and easy-to-use method that can handle many design parameters efficiently. These considerations have led to truss layout optimization, whose goal is to find the best material distribution within a given design domain discretized by a grid of nodal points and connected by tentative bars. <p><p>This general approach is well established for topology optimization where structural component sizes and system connectivity are simultaneously optimized. The range of applications covers limit analysis and identification of failure mechanisms in soils and masonries. However, to fully realize the potential of truss layout optimization for the design of lightweight structures, the consideration of geometrical variables is necessary. <p><p>The resulting truss geometry and topology optimization problem raises several fundamental and computational challenges. Our strategy to address the problem combines mathematical programming and structural mechanics: the structural properties of the optimal solution are used for devising the novel formulation. To avoid singularities arising in optimal configurations, the present approach disaggregates the equilibrium equations and fully integrates their basic elements within the optimization formulation. The resulting tool incorporates elastic and plastic design, stress and displacements constraints, as well as self-weight and multiple loading.<p><p>Besides, the inherent slenderness of lightweight structures requires the study of stability issues. As a remedy, we develop a conceptually simple but efficient method to include local and nodal stability constraints in the formulation. Several numerical examples illustrate the impact of stability considerations on the optimal design.<p><p>Finally, the investigation on realistic design problems confirms the practical applicability of the proposed method. It is shown how we can generate a range of optimal designs by varying design settings. In that regard, the computational design method mostly requires the designer a good knowledge of structural design to provide the initial guess. / Doctorat en Sciences de l'ingénieur / info:eu-repo/semantics/nonPublished