[pt] O Programa Nacional de Uso e Produção de Biodiesel destaca a produção de biodiesel a partir da mamona como alternativa energética não poluente e como gerador de empregos em regiões carentes. Todavia um empecilho à produção deste biodiesel advém da precariedade da produção da mamona, baseada em agricultores familiares pouco estruturados e com condições logísticas ruins. Assim, este trabalho visa a contribuir à resolução deste problema, procurando otimizar o planejamento estratégico desta cadeia de suprimentos de biodiesel em particular. O objetivo é minimizar os custos totais de transporte e de armazenagem de grãos dos produtores agrícolas às usinas de esmagamento. Uma importante peculiaridade deste problema é a incerteza da produção, que afeta o projeto da cadeia. Desta forma, foi proposto um Modelo de Programação Linear Inteira-Mista (PLIM) Estocástico, com formulação dois estágios e multi-estágio. Cabe ressaltar que este modelo foi testado em um caso real no semiárido brasileiro. Como resultado, são apresentadas as alocações de fluxos e entrepostos de custo mínimo para ambos os modelos. Por fim, é feita uma comparação entre estas formulações ressaltando que o ganho de flexibilidade obtido através do modelo multi-estágio se traduz em um menor custo logístico total. / [en] The Brazilian Program for Biodiesel Use and Production highlights the production of biodiesel from castor seeds, as a non-polluting energy source and as a job generation in poor regions. However, an obstacle to biodiesel production comes from the castor seeds poor production conditions and lack of infrastructure and logistics. This work aims to present a contribution to solve this problem, performing a strategic planning optimization of this biodiesel supply chain. The main objective is to minimize total storage costs and grains transportation to the crushing plant. An important peculiarity of this problem is the production uncertainty, which affects the supply chain design. Thus, we propose a stochastic Mixed Integer Linear Programming (MILP) model with two stage and multi-stage formulations. This model was tested on a real case in the Brazilian s semi-arid region. As a result, in order to obtain the minimum total cost, we present the logistics network flow design and warehouses assignments for both formulations. Lastly, we present a comparison between these formulations highlighting that the flexibility gain provided by the multi-stage model results in a lower total logistic cost.

[pt] Otimização de portfólio é uma técnica largamente utilizada para seleção de investimentos na área econômico-financeira. A primeira proposição neste sentido foi o modelo média- variância de Harry Markowitz, que utiliza, respectivamente, a média e a variância dos retornos do portfólio como medidas de retorno e de risco. Desde Markowitz muitas outras abordagens, que adotam medidas de risco alternativas, têm sido propostas, como por exemplo o modelo MiniMax, o modelo de desvio absoluto médio, a programação objetiva, o Value-at-Risk (VaR), o Conditional Value-at- Risk (CVaR) etc. Neste trabalho a idéia de otimização de portfólio é aplicada à área de comercialização de energia. O objetivo é apresentar abordagens para otimização de portfólio de contratos de energia, de modo a se definir a estratégia de comercialização de energia que maximize o valor esperado dos valores presentes das remunerações líquidas de uma empresa geradora, sujeito ao controle de sua exposição ao risco. São propostas três abordagens: a primeira adota a variância dos valores presentes das remunerações líquidas como medida de risco, a segunda adota o mínimo da distribuição como medida de risco e a terceira adota o CVaR como medida de risco. Em duas das três abordagens propostas, assume-se que os contratos candidatos a compor o portfólio são divididos em dois grupos: contratos de decisão imediata e possibilidades futuras de contratação. Com isto, a formulação do problema resulta em um modelo de otimização estocástica de dois estágios, que é resolvido via programação dinâmica dual estocástica. Resultados numéricos para o sistema elétrico brasileiro são apresentados e discutidos. / [en] Portfolio optimization has been widely used to select investments in the financial area. The first proposal in this topic was the Markowitz mean-variance approach, which uses, respectively, the mean and the variance as measures of portfolio return and risk. Since Markowitz many other approaches, which adopt alternative risk measures, have been proposed, e.g. the MiniMax model, the Mean Absolute Deviation model, the Goal Programming, the Value-at-Risk (VaR) and the Conditional Value-at-Risk (CVaR) etc. In this work the idea of portfolio optimization is applied to the energy commercialization area. The objective is to present approaches to portfolio optimization of energy contracts in order to determine the energy commercialization strategy that maximizes the expected present value of the cash flow of a generating company subject to the control of its risk exposure. Three approaches are proposed: the first adopts the variance of the present values as risk measure, the second adopts the minimum present value as risk measure and the third adopts the CVaR as risk measure. In the second and in the third approaches are assumed that the candidate contracts are divided into two sets: those of immediate decision and those that can be contracted in the future. This modeling leads to a large-scale two-stage stochastic programming problem that is solved by stochastic dual dynamic programming. Numerical results for the Brazilian power system are presented and discussed.

[pt] Todos os investidores, indivíduos e instituições, possuem obrigações e objetivos financeiros futuros. Por esse motivo, devem tomar decisões de investimento que sirvam a tais propósitos, considerando os riscos a que estão sujeitos. Com a finalidade de auxiliar o processo decisório, pode-se lançar mão de políticas de investimento ótimo, como a Gerência de Ativos e Passivos (Asset and Liability Management - ALM), objeto do presente estudo. O ALM é uma forma de combinar os ativos e passivos dos investidores, buscando alcançar as suas finalidades em termos financeiros. No que se refere aos investidores individuais, tema abordado neste trabalho, os supracitados objetivos podem corresponder, por exemplo, à aposentadoria almejada, bem como aos gastos com a educação dos filhos. Sendo assim, o presente estudo propõe apresentar uma metodologia de otimização sob incerteza, por meio da utilização de programação estocástica e técnicas de otimização de portfolio, aplicadas ao problema de gerenciamento de ativos e passivos de um investidor individual. O estudo tem como enfoque um modelo de programação linear multiperíodo, desenvolvido por Consiglio, Cocco e Zenios (2002), o qual maximiza a riqueza esperada do investidor no final do horizonte de planejamento, dado o nível de tolerância ao risco do indivíduo. Esse modelo será validado através da variação dos níveis de aversão ao risco do investidor, dos horizontes de planejamento e do retorno alvo desejado pelo investidor para ser alcançado no período final. / [en] All investors, individuals and institutions, have obligations and financial future goals. For this reason, they should make investment decisions that serve this purpose considering the risks they face. To assist in making decisions, it is possible to use the optimal investment policies, as the Asset and Liability Management, object of this work. The ALM, as is known, is a way to combine the assets and liabilities of investors seeking to achieve their goals in financial terms. In the case of individuals investors these goals can be seen as the individual s retirement and children s tuition. The present work proposes a methodology for optimization under uncertainty, employing both stochastic programming and portfolio optimization techniques, applied to the problem of managing assets and liabilities for an individual investor. The study is focused on a multi-period linear programming model developed by Consiglio, Cocco and Zenios (2002), which maximizes the expected wealth of the investor at the end of the planning horizon, given the individual s risk tolerance level. This model will be validated through the variation of the risk aversion level, the planning horizons and the target return that should be achieved on the final period.

The diploma thesis deals with penalty approach to stochastic optimization with chance constraints which are applied to structural mechanics. The problem of optimal design of beam dimensions is modeled and solved. The uncertainty is involved in the form of random load. The corresponding mathematical model contains a condition in the form of ordinary differencial equation that is solved by finite element method. The probability condition is approximated by several types of penalty functions. The results are obtained by computations in the MATLAB software.

La généralisation concomitante de consommateurs d'électricité flexibles et de producteurs imparfaitement contrôlables invite à utiliser les complémentarités de ces acteurs afin d'améliorer leur intégration dans les systèmes d'énergie. Dans le cadre de ces travaux de doctorat, la collaboration entre une flotte de véhicules électriques et une centrale photovoltaïque est étudiée. Un problème générique est tout d'abord défini afin d'augmenter la prévisibilité des échanges entre un réseau électrique et le système collaboratif ainsi créé qui devra respecter un profil d'engagement de puissance échangée. La gestion de ce système est traduite en un problème d'optimisation dans lequel on cherche à compenser les erreurs de prévision de la production photovoltaïque à l'aide de la flexibilité des recharges. Ce problème est multi-temporel du fait de la présence de batteries, stochastique à cause de la disponibilité des véhicules et des erreurs de prévision, et enfin de grande dimension puisqu'à l'échelle d'une flotte entière.Pour le résoudre, la modélisation du comportement et du vieillissement des batteries Li-ion est discutée afin d'établir des compromis entre justesse du modèle, impact sur la décision finale et coût de calcul. Par ailleurs, un modèle de Markov caché original est spécifiquement développé afin de capturer les structures temporelles de l'erreur de prévision de production photovoltaïque. Cette étude est fondée sur des données réelles de production d'une centrale et des données de prévision correspondantes.Le problème de recharge optimale d'une flotte de véhicules agrégée en une batterie équivalente est résolu par la méthode de la programmation dynamique stochastique. La sensibilité des lois de gestion obtenues est discutée vis à vis des modèles utilisés pour décrire l'erreur de prévision ou le comportement des batteries. Le vieillissement des batteries est traduit par plusieurs modèles, dont on examine les conséquences sur le dimensionnement optimal de la flotte de véhicules par rapport à la puissance crête de la centrale photovoltaïque.Enfin la puissance de recharge optimale pour chacun des véhicules de la flotte est déduite à l'aide d'un problème de partage qui est résolu par optimisation distribuée --- Alternating Direction Method of Multipliers --- et programmation dynamique. Une attention particulière est prêtée à la manière dont les préférences individuelles de chaque utilisateur peuvent être prises en compte au sein d'une flotte. Le cas d'une limitation des échanges d'information possibles entre les véhicules est investigué. Le dimensionnement optimal entre une flotte et une centrale photovoltaïque est finalement analysé pour plusieurs modèles économiques envisageables. L'interaction entre dimensionnement et gestion est traitée à l'aide d'une co-optimisation. / Simultaneous development of flexible electricity consumers and of intermittent renewable producers calls for using their complementarities. It could foster their overall integration in power systems. For the purpose of this doctoral thesis, the collaboration between an electric vehicle fleet and a photovoltaic plant is studied. First of all, a generic problem is set up to improve the predictability of the power exchange between the power grid and the so called collaboratif system. It should therefore fulfill a commitment profile constraint. The intraday management of this system consists in an optimisation problem which objective is to mitigate the production forecast errors by charging power flexibility. This is a multitime step problem, because of the battery intertia. The random availibility of vehicles and the forecast errors also make it stochastic. Finally there is a huge number of variables as it is spread other an entiere fleet.Upstream of the problem resolution, the modeling of the dynamic behaviour and of the aging of Lithium Ion batteries is discussed. It results in a range of compromises between precision, impact on the final decision and computational cost. Furthermore, a hidden Markov model is proposed and developped so as to handle temporal structures of the forecast error of the photovoltaic production. This analysis is based on production data of a real plant and on associated forecasts.An electric vehicle fleet is considered as an equivalent agregated battery. Its optimal charging power is sorted out using stochastic dynamic programming. The sensitivity of the resulting management strategies is assessed against the models which describe the production forecast error or battery behaviour. The battery aging is rendered by several models which we discuss the consequences over the optimal sizing of an electric vehicle fleet regarding to the plant power.Then the optimal charing power for each one of the vehicles among a fleet is deduced using a sharing problem. The resolution is carried out using distributed optimisation --- Alternating Direction Method of Multipliers --- and dynamic programming. A specific attention is devoted to the individual mobility priorities of the vehicles users. The vehicle charging power is thus differenticiated according to each one preferences. We also investigate a situation where information exchanges are limited. The optimal sizing of an electric vehicle fleet associated with a photovoltaic plant is finaly considered under several possibilities of economic model. The coupling between sizing and daily management is tackled thanks to a co-optimization.

abstract: Breeding seeds to include desirable traits (increased yield, drought/temperature resistance, etc.) is a growing and important method of establishing food security. However, besides breeder intuition, few decision-making tools exist that can provide the breeders with credible evidence to make decisions on which seeds to progress to further stages of development. This thesis attempts to create a chance-constrained knapsack optimization model, which the breeder can use to make better decisions about seed progression and help reduce the levels of risk in their selections. The model’s objective is to select seed varieties out of a larger pool of varieties and maximize the average yield of the “knapsack” based on meeting some risk criteria. Two models are created for different cases. First is the risk reduction model which seeks to reduce the risk of getting a bad yield but still maximize the total yield. The second model considers the possibility of adverse environmental effects and seeks to mitigate the negative effects it could have on the total yield. In practice, breeders can use these models to better quantify uncertainty in selecting seed varieties / Dissertation/Thesis / Masters Thesis Industrial Engineering 2019

No description available.

No description available.

No description available.

With the world becoming more conscious about achieving 1.5-degree scenario as promisedby the most powerful economies of the world, much needed push was received by the renewable energy technology providers.  This has led to an increased a share of energy production from renewables and a decrease in the fossil-based energy production with the overall energy production.  As a result, a large share of inertia of the system is lost and a big challenge in the name of flexibility is presented to the world of energy.  Virtual Power Plant is quite a novel and new concept to address the new generation challenge of flexibility and can offer various other benefits like competitivity,reliability, accessibility etc.  In this thesis, a commercial virtual power plant is studied by developing a mixed integer linear model to emulate the trading for short term markets with the risk mea- sures in a Nordic Electricity Framework.  Further, the developed model is implemented in a quite a new mathematical programming language known as “Julia”.  The model is implemented using a hypothetical portfolio consisting of a dispatchable unit, a battery system and a wind farm in the SE3 bidding zone of Sweden.  An investigation on varia- tion of imbalance costs in three different modes also has been carried out, to demonstratethe advantage of such a virtual power plant concept in reducing the imbalance costs. / För att uppfylla 1,5-gradersmålet som beslutats av världens ledande ekonomier har olika‌typer av förnybar energiproduktion fått ett stort uppsving. Detta har lett till ökad energiproduktion från förnybara källor och minskad energiproduktion från fossila källor. För elsystemen innebär en högre andel förnybar produktion minskad svängmassa ochökat behov av flexibilitet för att kompensera för variationen hos förnybara energikällor. Virtuella kraftverk är ett nytt koncept för att tillgodose behovet av flexibilitet och kanäven ge andra fördelar som konkurrenskraft och tillförlitlighet. I denna uppsats studeras ett virtuellt kraftverk genom att utveckla en optimeringsmodell för att emulera handeln i elmarknader med riskmått inom ett ramverk för den nordiska elmarknaden. Modellen implementeras i det nya programmeringsspråket Julia. Modellen innehåller en hypotetisk blandning av resurser bestående av ett planerbart kraftverk, ett batterisystem och en vindpark i elområdet SE3 i Sverige. Balanseringskostnaderna i tre olika modeller undersöks för att visa potentialen hos det virtuella kraftverket att minska dessa kostnader.