This thesis grew out of a problem encountered by a subsidiary of a Swedish multinational industrial corporation. This subsidiary is responsible for the corporation’s customer financing activities. In the thesis, we refer to these entities as the Division and the Corporation. The Division needed to find a new approach to finance its customer loan portfolio. Risk control and return maximization were important aspects of this need. The objective of this thesis is to devise and implement a method that allows the Division to make optimal funding decisions, given a certain risk limit. We propose a funding approach based on stochastic programming. Our approach allows the Division’s portfolio manager to minimize the funding costs while hedging against market risk. We employ principal component analysis and Monte Carlo simulation to develop a multicurrency scenario generation model for interest and exchange rates. Market rate scenarios are used as input to three different optimization models. Each of the optimization models presents the optimal funding decision as positions in a unique set of financial instruments. By choosing between the optimization models, the portfolio manager can decide which financial instruments he wants to use to fund the loan portfolio. To validate our models, we perform empirical tests on historical market data. Our results show that our optimization models have the potential to deliver sound and profitable funding decisions. In particular, we conclude that the utilization of one of our optimization models would have resulted in an increase in the Division’s net income over the past 3.5 years.

The Swedish National Debt Office (SNDO) is the Swedish Government’s financial administration. It has several tasks and the main one is to manage the central government’s debt in a way that minimizes the cost with due regard to risk. The debt management problem is to choose currency composition and maturity profile - a problem made difficult because of the many stochastic factors involved. The SNDO has created a simulation model to quantitatively analyze different aspects of this problem by evaluating a set of static strategies in a great number of simulated futures. This approach has a number of drawbacks, which might be handled by using a financial optimization approach based on Stochastic Programming. The objective of this master’s thesis is thus to apply financial optimization on the Swedish government’s strategic debt management problem, using the SNDO’s simulation model to generate scenarios, and to evaluate this approach against a set of static strategies in fictitious future macroeconomic developments. In this report we describe how the SNDO’s simulation model is used along with a clustering algorithm to form future scenarios, which are then used by an optimization model to find an optimal decision regarding the debt management problem. Results of the evaluations show that our optimization approach is expected to have a lower average annual real cost, but with somewhat higher risk, than a set of static comparison strategies in a simulated future. These evaluation results are based on a risk preference set by ourselves, since the government has not expressed its risk preference quantitatively. We also conclude that financial optimization is applicable on the government debt management problem, although some work remains before the method can be incorporated into the strategic work of the SNDO.

Nowadays, power systems are dealing with some new challenges raisedby the major changes that have been taken place since 80’s, e.g., deregu-lation in electricity markets, significant increase of electricity demands andmore recently large-scale integration of renewable energy resources such aswind power. Therefore, system operators must make some adjustments toaccommodate these changes into the future of power systems.One of the main challenges is maintaining the system stability since theextra stress caused by the above changes reduces the stability margin, andmay lead to rise of many undesirable phenomena. The other important chal-lenge is to cope with uncertainty and variability of renewable energy sourceswhich make power systems to become more stochastic in nature, and lesscontrollable.Flexible AC Transmission Systems (FACTS) have emerged as a solutionto help power systems with these new challenges. This thesis aims to ap-propriately utilize such devices in order to increase the transmission capacityand flexibility, improve the dynamic behavior of power systems and integratemore renewable energy into the system. To this end, the most appropriatelocations and settings of these controllable devices need to be determined.This thesis mainly looks at (i) rotor angle stability, i.e., small signal andtransient stability (ii) system operation under wind uncertainty. In the firstpart of this thesis, trajectory sensitivity analysis is used to determine themost suitable placement of FACTS devices for improving rotor angle sta-bility, while in the second part, optimal settings of such devices are foundto maximize the level of wind power integration. As a general conclusion,it was demonstrated that FACTS devices, installed in proper locations andtuned appropriately, are effective means to enhance the system stability andto handle wind uncertainty.The last objective of this thesis work is to propose an efficient solutionapproach based on Benders’ decomposition to solve a network-constrained acunit commitment problem in a wind-integrated power system. The numericalresults show validity, accuracy and efficiency of the proposed approach. / <p>The Doctoral Degrees issued upon completion of the programme are issued by Comillas Pontifical University, Delft University of Technology and KTH Royal Institute of Technology. The invested degrees are official in Spain, the Netherlands and Sweden, respectively.QC 20141028</p>

Operation of the electricity markets is subject to a number of strict and specific constraints such as continuous load-generation balance, security of supply, and generation technology related limitations. Contributions have been made to two important problems of the Electricity Markets, in the context of this study. In this study, Price Based Unit Commitment problem in the literature, which is a tool for the GENCO for operations planning, is extended considering the interdependencies between the Natural Gas (NG) and Electricity infrastructures and the uncertainty of Wind Power generation. The effect of the NG infrastructure physical limitations is considered via linearized NG transmission system equations, and the Wind energy sources and conventional generation resource uncertainties are simulated by Monte-Carlo simulations. The contribution of the forward energy Bilateral Contracts (BC), as a financial risk hedging tool is also included by modeling these in the proposed PBUC framework. In the case studies , it is observed that a GENCO could prevent its financial losses due to NG interruptions, by depositing only a portion of the midterm interrupted NG in the storage facilities. The Security Constrained Unit Commitment (SCUC) Problem is widely accepted tool in the industry which models the market clearing process. This study integrates two novelties to the SCUC problem / &bull / A discrete demand response model to consider active participation of the consumers, &bull / A hybrid deterministic/stochastic contingency model to represent the N-1 contingencies together with the uncertainties related with the wind power generation and system load. It is observed that the curtailment of available wind power capacity would enable the TSO to take corrective actions against occurrence of the contingencies and realization of the uncertainties in the most possible economical manner.

Modern aerospace systems rely heavily on legacy platforms and their derivatives. Historical examples show that after a vehicle design is frozen and delivered to a customer, successive upgrades are often made to fulfill changing requirements. Current practices of adapting to emerging needs with derivative designs, retrofits, and upgrades are often reactive and ad-hoc, resulting in performance and cost penalties. Recent DoD acquisition policies have addressed this problem by establishing a general paradigm for design for lifelong evolution. However, there is a need for a unified, practical design approach that considers the lifetime evolution of an aircraft concept by incorporating future requirements and technologies. This research proposes a systematic approach with which the decision makers can evaluate the value and risk of a new aircraft development program, including potential derivative development opportunities. The proposed Evaluation of Lifelong Vehicle Evolution (EvoLVE) method is a two- or multi-stage representation of the aircraft design process that accommodates initial development phases as well as follow-on phases. One of the key elements of this method is the Stochastic Programming with Recourse (SPR) technique, which accounts for uncertainties associated with future requirements. The remedial approach of SPR in its two distinctive problem-solving steps is well suited to aircraft design problems where derivatives, retrofits, and upgrades have been used to fix designs that were once but no longer optimal. The solution approach of SPR is complemented by the Risk-Averse Strategy Selection (RASS) technique to gauge risk associated with vehicle evolution options. In the absence of a full description of the random space, a scenario-based approach captures the randomness with a few probable scenarios and reveals implications of different future events. Last, an interactive framework for decision-making support allows simultaneous navigation of the current and future design space with a greater degree of freedom. A cantilevered beam design problem was set up and solved using the SPR technique to showcase its application to an engineering design setting. The full EvoLVE method was conducted on a notional multi-role fighter based on the F/A-18 Hornet.

Submitted by YURI PRADO DE SOUZA (yuriprado.uff@gmail.com) on 2018-10-17T22:40:46Z No. of bitstreams: 1 Dissertação v60 - final.pdf: 1880394 bytes, checksum: 1c4ca28a4089a492a49b54e291c33dea (MD5) / Rejected by Pamella Benevides Gonçalves null (pamella@feg.unesp.br), reason: Solicitamos que realize correções na submissão seguindo as orientações abaixo: Rever a ordenação dos elementos pré-textuais ... capa, folha de rosto ... ficha catalográfica ... • A capa e ficha catalográfica não são consideradas para contagem de páginas. a paginação deve aparecer no canto superior direito a partir da introdução, realizei a contagem das páginas e seu trabalho deve com o número (14)*, após você precisa atualizar a numeração na ficha catalográfica, nas listas e no sumário. • Resumo: Apenas palavra Resumo e Abstract devem ser centralizada; o resumo deve ser em parágrafo único. (favor ver exemplo no template ou diretrizes) o As palavras-chave e keyword devem ser separadas entre si por ponto final e também finalizadas por ponto. (favor ver exemplo no template ou diretrizes) • A lista de figuras existem algumas que não aparece o título, a numeração das figuras devem ser continuas independente do capitulo. • Sumário: deve ter os mesmo destaques tipográfico que as seções do trabalho, deve ser alinhado à esquerda (veja exemplo no template ou diretrizes) • Favor revisar as todos os indicativos de seção em seu trabalho e no sumário • INDICATIVO DE SEÇÃO Os títulos das seções devem começar na parte superior da folha e separados do texto que os sucede por um espaço de 1,5 entrelinhas. Da mesma forma, os títulos das subseções devem ser separados do texto que os precede e que os sucede por por um espaço de 1,5 entrelinhas. Os títulos das seções devem ser destacados tipograficamente, da primária a quinária. As seções primárias por serem as principais divisões de texto, devem iniciar em folha distinta, no final dos indicativos de seção não tem ponto final exemplo 7 MODELO DE REFERÊNCIA (seção primária) - caixa alta/negrito 7.1 PUBLICAÇÃO PERIÓDICA (seção secundária) - caixa alta sem negrito 7.1.1 Publicação periódica no todo (seção terciária) negrito 7.1.1.1 Artigo de periódico (seção quaternária) - sem negrito 7.1.1.1. Com autor pessoal (seção quinária) - Itálico e negrito • Qualquer que seja o tipo de ilustração (figuras, desenhos, gráficos, diagramas,fluxogramas, fotografias, mapa, planta, quadro, imagem entre outros) sua identificação (título) aparece na parte superior com letra tamanho 12; o Na parte inferior, Tamanho da letra 10, indicar a fonte consultada (elemento obrigatório, mesmo que seja produção do próprio autor), notas e outras informações necessárias à sua compreensão. o Devem conter a fonte mesmo que elaborada pelo autor. o Ex: Fonte: Autor Fonte: Autoria própria (favor ver exemplo no template ou diretrizes) • As fontes das ilustrações, tabelas e quadros não podem ser links . Areferência deve ser informada ao final, seguindo os padrões da ABNT.Para indicar a fonte, deve ser colocada a autoria e o ano entre parênteses. Ex.: Martins (2010). Quando uma referência for retirada de um meio eletrônico deve-se identificar uma autoria para o que é visualizado na página; se não houver título, escrever uma pequena descrição do que foi visto e seguir com os dados: disponível em:<endereço eletronico> . Acesso em: xx mes xxxx. A autoria pode ser uma pessoa física, uma Instituição, uma empresa, uma pessoa jurídica e até o nome do próprio site. Ex.: ECOVILAS. Condomínios autossustentados e permaculturais. Disponível em: <http://www.ecoovilas.com/projetos/permacultura>. Acesso em: 10 out. 2017. Será colocado na Fonte: Ecovilas (2017) • Referências. A palavra Referências deve ser centralizada, e não conter numeração de seção; As referencias devem ser justificadas, espaço simples com um espaço simples(enter) entre elas. • Sobre a elaboração das referencias e citações e formatação favor solicitar ajuda com URGÊNCIA a bibliotecária Juciene (juciene.pedroso@unesp.br) Mais informações acesse o link: http://www2.feg.unesp.br/Home/Biblioteca21/diretrizes-2016.pdf Agradecemos a compreensão. on 2018-10-18T12:54:42Z (GMT) / Submitted by YURI PRADO DE SOUZA (yuriprado.uff@gmail.com) on 2018-10-19T18:53:48Z No. of bitstreams: 2 Dissertação v60 - final.pdf: 1880394 bytes, checksum: 1c4ca28a4089a492a49b54e291c33dea (MD5) Dissertação v-61 formatado2.pdf: 1810118 bytes, checksum: 4638b9426aac62a064b565b38ffda481 (MD5) / Approved for entry into archive by Pamella Benevides Gonçalves null (pamella@feg.unesp.br) on 2018-10-19T19:04:38Z (GMT) No. of bitstreams: 1 souza_yp_me_guara.pdf: 1810118 bytes, checksum: 4638b9426aac62a064b565b38ffda481 (MD5) / Made available in DSpace on 2018-10-19T19:04:38Z (GMT). No. of bitstreams: 1 souza_yp_me_guara.pdf: 1810118 bytes, checksum: 4638b9426aac62a064b565b38ffda481 (MD5) Previous issue date: 2018-08-27 / Neste trabalho propõe-se uma abordagem para o Problema de Balanceamento de Linhas de Montagem (do inglês, Assembly Line Balancing Problem - ALBP) para aumentar a eficiência de uma indústria montadora de veículos. O ALBP caracteriza-se como um problema de sequenciamento de tarefas em estações de trabalho classificado como um problema de Otimização Combinatória NP-difícil e, portanto, a solução exata do problema em ambientes reais geralmente implica em elevado custo computacional. Para resolver o ALBP, foram formulados um modelo matemático de otimização inteira mista para obtenção de soluções determinísticas e um modelo estocástico com recurso que considera a incerteza dos tempos de execução das tarefas pelos operadores. A motivação para o desenvolvimento do presente trabalho decorre da observação de interrupções constantes do fluxo de produção nesta indústria, atribuídas às mais diversas naturezas, e que causavam transtornos e elevados níveis de estresse aos trabalhadores. Ambos os modelos, determinístico e estocástico, aumentaram a capacidade de produção de 196 unidades/dia para 245 e 233 unidades/dia, respectivamente. O modelo estocástico aumentou o tempo de ciclo CT em 5,6% quando comparado ao modelo determinístico, embora diminua a capacidade efetiva em 4,8% Porém, não considerar a incerteza no tempo de execução das tarefas pode diminuir a quantidade produzida em até 10,6%. Contrariamente ao entendimento comum em linhas de montagem, este trabalho conclui que reduzir os tempos de ociosidade aos níveis mínimos é prejudicial à produtividade de linhas de montagem. Isto se deve ao fato de que uma parcela do tempo atribuído à ociosidade dos operadores, na verdade contêm um tempo adicional gerado pela incerteza do tempo de execução das tarefas. Os resultados sugerem que a abordagem do ALBP sob incerteza contribui para o aumento dos índices de capacidade operacional da empresa. Devido ao grande esforço computacional necessário para a solução dos modelos de otimização propostos (determinístico e estocástico), não se consegue resolver, em um tempo computacional razoável, exemplares de dimensões reais do problema. Em vista disto, o trabalho propõe também uma heurística para a solução do ALBP visando minimizar o tempo de ciclo. Experimentos computacionais sugerem que a heurística proposta obtém resultados razoáveis para grandes exemplares do problema em um tempo computacional pequeno / This work proposes solution approaches to the Assembly Line Balancing Problem (ALBP) to increase the efficiency of a vehicle assembler industry. The ALBP is characterized as a task sequencing in workstations which is classified as a NP-hard Combinatorial Optimization problem and, therefore, the exact solution of the problem in real environments usually implies a high computational cost. In order to solve the ALBP, a mathematical model of mixed integer optimization to obtain deterministic solutions and a stochastic model with resource that considers the uncertainty of the execution times of the tasks by the operators were formulated. The motivation for the development of this work stems from the constant interruptions of the production flow in this industry, attributed to the most diverse natures, which cause disorders and high levels of stress to the workers. The deterministic and stochastic models increased the production capacity from 196 units / day to 245 and 233 units / day, respectively. The stochastic model increased the cycle time by 5.6% when compared to the deterministic model, although it reduced the effective capacity by 4.8%, which is equivalent to 12 vehicles / day. However, not considering the uncertainty in task execution times can decrease the amount produced by up to 10.6% or 26 vehicles / day. Contrary to the most acceptable idea, this work concludes that reducing idle times to minimum levels is detrimental to assembly line productivity. This is due to the fact that a portion of the time attributed to the idleness of the operators actually contains an additional time generated by the uncertainty of the execution time of the tasks. The results suggest that the approach of the ALBP under uncertainty contributes to the increase of the indices of operational capacity of the company. Due to the great computational effort required to solve the proposed optimization models (deterministic and stochastic), it is not possible to solve real instances of the problem in a reasonable computational time. In view of this, this work also proposes a heuristic for the ALBP solution in order to minimize the cycle time. Computational experiments suggest that the proposed heuristic obtains reasonable results for large instances of the problem in a small computational time

Chaque année, plus de 400 catastrophes naturelles frappent le monde. Pour aider les populations touchées, les organisations humanitaires stockent par avance de l’aide d’urgence dans des entrepôts. Cette thèse propose des outils d’aide à la décision pour les aider à localiser et dimensionner ces entrepôts. Notre approche repose sur la construction de scénarios représentatifs. Un scénario représente la survenue d’une catastrophe dont on connaît l’épicentre, la gravité et la probabilité d’occurrence. Cette étape repose sur l’exploitation et l’analyse de bases de données des catastrophes passées. La seconde étape porte sur la propagation géographique de la catastrophe et détermine son impact sur la population des territoires touchés. Cet impact est fonction de la vulnérabilité et de la résilience du territoire. La vulnérabilité mesure la valeur attendue des dégâts alors que la résilience estime la capacité à résister au choc et à se rétablir rapidement. Les deux sont largement déterminées par des facteurs économiques et sociaux, soit structurels (géographie, PIB…) ou politiques (existence d’infrastructure d’aide, normes de construction…). Nous proposons par le biais d’analyses en composantes principales (ACP) d’identifier les facteurs influents de résilience et de vulnérabilité, puis d’estimer le nombre de victimes touchées à partir de ces facteurs. Souvent, les infrastructures (eau, télécommunication, électricité, voies de communication) sont détruits ou endommagés par la catastrophe (ex : Haïti en 2010). La dernière étape a pour objectif d’évaluer les impacts logistiques en ce qui concerne : les restrictions des capacités de transport existant et la destruction de tout ou partie des stocks d’urgence. La suite de l’étude porte sur la localisation et le dimensionnement du réseau d’entrepôt. Nos modèles présentent l’originalité de tenir compte de la dégradation des ressources et infrastructures suite due à la catastrophe (dimension résilience) et de chercher à optimiser le rapport entre les coûts engagés et le résultat obtenu (dimension efficience). Nous considérons d’abord un scénario unique. Le problème est une extension d’un problème de location classique. Puis, nous considérons un ensemble de scénarios probabilisés. Cette approche est indispensable à la considération du caractère très incertain des catastrophes humanitaires. L’ensemble de ces contributions a été confronté à la réalité des faits dans le cadre d’une application au cas des crises récurrentes du Pérou. Ces crises, essentiellement dues aux tremblements de terre et aux inondations (El Niño), imposent la constitution d’un réseau logistique de premiers secours qui soit résilient et efficient. / Every year, more than 400 natural disasters hit the world. To assist those affected populations, humanitarian organizations store in advance emergency aid in warehouses. This PhD thesis provides tools for support decisions on localization and sizing of humanitarian warehouses. Our approach is based on the design of representative and realistic scenarios. A scenario expresses some disasters’ occurrences for which epicenters are known, as well as their gravity and frequency. This step is based on the exploitation and analysis of databases of past disasters. The second step tackles about possible disaster’s propagation. The objective consists in determining their impact on population on each affected area. This impact depends on vulnerability and resilience of the territory. Vulnerability measures expected damage values meanwhile resilience estimates the ability to withstand some shock and recover quickly. Both are largely determined by social and economic factors, being structural (geography, GDP, etc.) or political (establishment or not relief infrastructure, presence and strict enforcement of construction standards, etc.). We propose through Principal Component Analysis (PCA) to identify, for each territory, influential factors of resilience and vulnerability and then estimate the number of victims concerned using these factors. Often, infrastructure (water, telecommunications, electricity, communication channels) are destroyed or damaged by the disaster (e.g. Haiti in 2010). The last step aims to assess the disaster logistics impact, specifically those related to with: transportation flows capacity limitations and destruction of all or part of emergency relief inventories. The following of our study focuses on location and allocation of a warehouses’ network. The proposed models have the originality to consider potential resources and infrastructure degradation after a disaster (resilience dimension) and seek optimizing the equilibrium between costs and results (effectiveness dimension). Initially we consider a single scenario. The problem is an extension of classical location studies. Then we consider a set of probable scenarios. This approach is essential due to the highly uncertain character of humanitarian disasters. All of these contributions have been tested and validated through a real application case: Peruvian recurrent disasters. These crises, mainly due to earthquakes and floods (El Niño), require establishment of a first aid logistics network that should be resilient and efficient.

[pt] Os combustíveis de baixo impacto ambiental estão em destaque na mídia e na sociedade, atualmente. Neste contexto, o hidrogênio, fonte de energia limpa, tem um grande potencial. Entretanto, ainda não existe uma infraestrutura adequada para sua comercialização. O crescimento da demanda por hidrogênio é de difícil previsão, gerando um alto grau de incerteza na definição das necessidades de capacidades futuras de sua rede logística. Esta dissertação propõe uma metodologia para o planejamento do projeto da cadeia de suprimentos de hidrogênio para uso em transporte. Para representar o problema e avaliar diferentes alternativas de investimentos em infraestrutura logística foi desenvolvido um modelo matemático estocástico de dois estágios utilizando programação linear inteira mista (PLIM). O elevado nível de incerteza desta cadeia aumenta a complexidade do modelo, requerendo uma grande quantidade de cenários, inviabilizando sua otimização. Para contornar esta dificuldade, foi utilizada a técnica de aproximação por média amostral (SAA). Esta abordagem gera soluções, cuja qualidade pode ser estatisticamente avaliada utilizando-se um número reduzido de cenários. A metodologia proposta foi aplicada em um estudo de caso com dados reais da cadeia de suprimentos de hidrogênio líquido da Grã-Bretanha. Os gaps de otimalidade gerados nestes testes foram inferiores a 1 por cento, demonstrando a adequação do método desenvolvido. Mesmo com o alto nível de incerteza do problema, o SAA possibilitou definir como, quando, e onde investir. Os resultados obtidos devem contribuir para proporcionar avanços na criação de uma infraestrutura apropriada para a comercialização do hidrogênio. / [en] Nowadays, fuels with low environmental impact are highlighted in media and society. In this context, hydrogen, as a clean energy source, has a great potential. However, there is still no appropriate infrastructure for its commercialization. The prediction of demand for hydrogen is difficult, generating a high degree of uncertainty in the definition of capacity needs in the future for its logistics network. This work proposes a methodology for the design of the hydrogen supply chain for use in transportation. To represent the problem and evaluate alternatives to invest in logistics infrastructure, a two-stage stochastic mixed-integer programming was developed. The high degree of uncertainty in this chain increases the complexity of the mathematical model, requiring a huge number of scenarios which makes its optimization impossible. To overcome this difficulty, the technique of sample average approximation (SAA) is used. This approach generates solutions, whose quality can be statistically evaluated using a reduced number of scenarios. The proposed methodology was tested in a study case with real data from Great Britain s liquid hydrogen supply chain. The optimal gaps generated in these tests were below 1 percent, demonstrating the adequacy of the developed methodology. Even with the high level of uncertainty of the problem, the propose methodology using SAA technique can define how, when, and where to invest. The results should be helpful in advancing the creation of an appropriate infrastructure for hydrogen commercialization.

[pt] Na indústria do petróleo e gás, uma das atividades de maior importância é a intervenção em poços para serviços de manutenção, a qual é necessária para garantir a produção de petróleo. Estas intervenções são realizadas por sondas workover que são disponibilizadas para atender uma grande quantidade de poços segundo um itinerário. Nesta tese são propostos três modelos de programação linear inteira para abordar eficientemente o problema de intervenção em poços terrestres de petróleo. O primeiro modelo determina o itinerário de um conjunto de sondas homogêneas, visando minimizar a perda total de produção. Este modelo é um aprimoramento do modelo proposto por Costa e Ferreira Filho (2004). O segundo modelo é uma extensão do anterior e considera também o dimensionamento de uma frota de sondas heterogênea, procurando minimizar o custo de perda de produção e o custo de aluguel de sondas. O terceiro modelo é uma abordagem estocástica que estende o segundo modelo e consiste em dimensionar uma frota de sondas considerando o tempo de intervenção incerto. A incerteza do tempo de intervenção é representada mediante a geração de cenários, usando para este fim os métodos de Monte Carlo, Redução de Cenários e Quasi-Monte Carlo. Os testes de estabilidade propostos por Kaut e Wallace (2003) são aplicados para avaliar os métodos de geração de cenários e estabelecer o número de cenários adequados para resolver o problema. Para avaliar o desempenho dos modelos propostos, diversos experimentos computacionais foram realizados em instâncias de pequeno, médio e grande porte. Todas as instâncias são baseadas em casos reais no Brasil. Os resultados mostram que os modelos propostos foram capazes de resolver todas as instâncias utilizadas, inclusive aquelas de grande porte, demonstrando serem eficientes quando comparadas com várias metaheurísticas, pois produzem soluções exatas em um curto tempo computacional. Uma análise do impacto nas soluções quando ocorre uma mudança no preço de petróleo e no horizonte de planejamento também é realizada. A metodologia de resolução empregada no terceiro modelo mostrou que o método Quasi-Monte Carlo proporcionou os melhores cenários para representar a incerteza e também o potencial do modelo para resolver problemas de grande porte. / [en] In the oil and gas industry, one of the most important activities is the intervention in wells for maintenance services, which is necessary to ensure the production of oil. These interventions are performed by workover rigs that are available to serve a large number of wells according to a schedule. In this thesis, we proposed three integer linear programming models to efficiently address the problem of intervention in onshore oil wells. The first model determines the schedule of a set of homogeneous rigs, with the objective of minimizing the total production loss. This model is an improvement of the model proposed by Costa and Ferreira Filho (2004). The second model is an extension of the previous one and also considers the sizing of a heterogeneous rig fleet, with the objective of minimizing the production loss cost and the rig rental cost. The third model is a stochastic approach that extends the second model and consists of sizing a rig fleet considering the uncertainty in the intervention time. The uncertainty in the intervention time is represented by the generation of scenarios, using for this purpose the Monte Carlo, Scenario Reduction, and Quasi-Monte Carlo methods. The stability tests proposed by Kaut and Wallace (2003) are applied to evaluate the scenario generation methods and to establish the number of appropriate scenarios to solve the problem. To evaluate the performance of the proposed models, several computational experiments were performed in small, medium and large instances. All instances are based on real cases in Brazil. The results show that the proposed models were able to solve all of the instances considered, including the large instances, proving to be efficient when compared to various metaheuristics, as they produce exact solutions in small computational time. An analysis of the impact on the solutions when there is a change in the oil price and the planning horizon is also carried out. The resolution methodology employed in the third model showed that the Quasi-Monte Carlo method provided the best scenarios to represent the uncertainty and also the potential of the model to solve large-scale problems.

[pt] Apesar de existir na literatura modelos propostos para gestão de estoques, as premissas consideradas por tais modelos podem inviabilizar suas aplicações. Este trabalho propõe uma metodologia de programação estocástica para reposição e controle de estoques de produto único numa rede logística de duas camadas. O enfoque revisão periódica proposto pode considerar tanto atendimentos à demanda em atraso (backorders) como vendas perdidas (lost sales) sem restrição de pedidos pendentes. Além disso, a fim de alcançar um melhor nível de serviço para o cliente, é introduzida uma regra de rateio proporcional a quantidade faltante do item em estoque no centro de distribuição para atender simultaneamente a demanda de todos os varejistas, a qual é capaz de lidar com as alocações negativas da falta. A periodicidade e o nível alvo da posição dos estoques são determinados através de modelos de programação estocástica de dois estágios e de uma técnica baseada em simulação de Monte Carlo, conhecida como Sample Average Approximation, que levam em conta a natureza incerta dos níveis de demanda pelo item por meio da geração de conjuntos finitos de cenários. Os equivalentes determinísticos são apresentados como modelos de programação não-linear inteira mista e em seguida linearizados. Experimentos numéricos com a metodologia proposta para instâncias do problema geradas aleatoriamente demonstram seu potencial ao obter resultados com erros de aproximadamente 1 por cento. / [en] Although several methods for inventory management are proposed in the literature, the required assumptions can hinder their application in practice. This work proposes a methodology for stock replenishment in two-echelon logistic networks through stochastic programming, considering a single item, periodic review and uncertain demands. The proposed approach is flexible enough to consider backlogs and lost sales cases without limitations on the number of outstanding orders. Also, in order to achieve better customer service, we introduce a variable rationing rule for quantities of the item in short at the distribution center to meet simultaneously all the demands of the retailers, dealing with imbalances or negative allocations of quantities of the item in short. The optimal review periodicity and the target level for inventory position are determined through two-stage stochastic programming models and a Monte Carlo simulation based-technique, known as Sample Average Approximation, which takes into account the uncertain nature of the item demand levels through the generation of finite sets of scenarios. The deterministic equivalent models are presented as mixed-integer non-linear programming models, which are then linearized. Numerical experiments with the proposed approach for instances of the problem randomly generated shows its potential, as the errors of the obtained results are around 1 percent.