The design exploration method for adaptive design systems is developed to facilitate the pursuit of a balance between the efficiency and accuracy in systems engineering design. The proposed method is modified from an existing multiscale material robust design method, the Inductive Design Exploration Method (IDEM). The IDEM is effective in managing uncertainty propagation in the model chain. However, it is not an appropriate method in other systems engineering design outside of original design domain due to its high computational cost. In this thesis, the IDEM is augmented with more efficient solution search methods to improve its capability for efficiently exploring robust design solutions in systems engineering design. The accuracy of the meta-model in engineering design is one uncertainty source. In current engineering design, response surface model is widely used. However, this method is shown as inaccurate in fitting nonlinear models. In this thesis, the local regression method is introduced as an alternative of meta-modeling technique to reduce the computational cost of simulation models. It is proposed as an appropriate method in systems design with nonlinear simulations models. The proposed methods are tested and verified by application to a Multifunctional Energetic Materials design and a Photonic Crystal Coupler and Waveguide design. The methods are demonstrated through the better accuracy of the local regression model in comparison to the response surface model and the better efficiency of the design exploration method for adaptive design systems in comparison to the IDEM. The proposed methods are validated theoretically and empirically through application of the validation square.

Thesis (Ph. D.)--Aerospace Engineering, Georgia Institute of Technology, 2008. / Committee Chair: Dr. Mavris, Dimitri; Committee Member: Dr. Hollingsworth, Peter; Committee Member: Dr. McMichael, Jim; Committee Member: Dr. Saleh, Joseph; Committee Member: Dr. Schrage, Daniel.

Thesis (Ph.D.)--Aerospace Engineering, Georgia Institute of Technology, 2009. / Committee Chair: Mavris, Dimitri; Committee Member: Bishop, Carlee; Committee Member: McMichael, James; Committee Member: Nixon, Janel; Committee Member: Schrage, Daniel

Machine learning algorithms are invented to learn from data and to use data to perform predictions and analyses. Many agencies are now using machine learning algorithms to present services and to perform tasks that used to be done by humans. These services and tasks include making high-stake decisions. Determining the right decision strongly relies on the correctness of the input data. This fact provides a tempting incentive for criminals to try to deceive machine learning algorithms by manipulating the data that is fed to the algorithms. And yet, traditional machine learning algorithms are not designed to be safe when confronting unexpected inputs. In this dissertation, we address the problem of adversarial machine learning; i.e., our goal is to build safe machine learning algorithms that are robust in the presence of noisy or adversarially manipulated data. Many complex questions -- to which a machine learning system must respond -- have complex answers. Such outputs of the machine learning algorithm can have some internal structure, with exponentially many possible values. Adversarial machine learning will be more challenging when the output that we want to predict has a complex structure itself. In this dissertation, a significant focus is on adversarial machine learning for predicting structured outputs. In this thesis, first, we develop a new algorithm that reliably performs collective classification: It jointly assigns labels to the nodes of graphed data. It is robust to malicious changes that an adversary can make in the properties of the different nodes of the graph. The learning method is highly efficient and is formulated as a convex quadratic program. Empirical evaluations confirm that this technique not only secures the prediction algorithm in the presence of an adversary, but it also generalizes to future inputs better, even if there is no adversary. While our robust collective classification method is efficient, it is not applicable to generic structured prediction problems. Next, we investigate the problem of parameter learning for robust, structured prediction models. This method constructs regularization functions based on the limitations of the adversary in altering the feature space of the structured prediction algorithm. The proposed regularization techniques secure the algorithm against adversarial data changes, with little additional computational cost. In this dissertation, we prove that robustness to adversarial manipulation of data is equivalent to some regularization for large-margin structured prediction, and vice versa. This confirms some of the previous results for simpler problems. As a matter of fact, an ordinary adversary regularly either does not have enough computational power to design the ultimate optimal attack, or it does not have sufficient information about the learner's model to do so. Therefore, it often tries to apply many random changes to the input in a hope of making a breakthrough. This fact implies that if we minimize the expected loss function under adversarial noise, we will obtain robustness against mediocre adversaries. Dropout training resembles such a noise injection scenario. Dropout training was initially proposed as a regularization technique for neural networks. The procedure is simple: At each iteration of training, randomly selected features are set to zero. We derive a regularization method for large-margin parameter learning based on dropout. Our method calculates the expected loss function under all possible dropout values. This method results in a simple objective function that is efficient to optimize. We extend dropout regularization to non-linear kernels in several different directions. We define the concept of dropout for input space, feature space, and input dimensions, and we introduce methods for approximate marginalization over feature space, even if the feature space is infinite-dimensional. Empirical evaluations show that our techniques consistently outperform the baselines on different datasets.

La qualité d'un plan d'exécution engendré par un optimiseur de requêtes est fortement dépendante de la qualité des estimations produites par le modèle de coûts. Malheureusement, ces estimations sont souvent imprécises. De nombreux travaux ont été menés pour améliorer la précision des estimations. Cependant, obtenir des estimations précises reste très difficile car ceci nécessite une connaissance préalable et détaillée des propriétés des données et des caractéristiques de l'environnement d'exécution. Motivé par ce problème, deux approches principales de méthodes d'optimisation ont été proposées. Une première approche s'appuie sur des valeurs singulières d'estimations pour choisir un plan d'exécution optimal. A l'exécution, des statistiques sont collectées et comparées à celles estimées. En cas d'erreur d'estimation, une ré-optimisation est déclenchée pour le reste du plan. A chaque invocation, l'optimiseur associe des valeurs spécifiques aux paramètres nécessaires aux calculs des coûts. Cette approche peut ainsi induire plusieurs ré-optimisations d'un plan, engendrant ainsi de mauvaises performances. Dans l'objectif d'éviter cela, une approche alternative considère la possibilité d'erreurs d'estimation dès la phase d'optimisation. Ceci est modélisé par l'utilisation d'un ensemble de points d'estimations pour chaque paramètre présumé incertain. L'objectif est d'anticiper la réaction à une sous-optimalité éventuelle d'un plan d'exécution. Les méthodes dans cette approche cherchent à générer des plans robustes dans le sens où ils sont capables de fournir des performances acceptables et stables pour plusieurs conditions d'exécution. Ces méthodes supposent souvent qu'il est possible de trouver un plan robuste pour l'ensemble de points d'estimations considéré. Cette hypothèse reste injustifiée, notamment lorsque cet ensemble est important. De plus, la majorité de ces méthodes maintiennent sans modification un plan d'exécution jusqu'à la terminaison. Cela peut conduire à de mauvaises performances en cas de violation de la robustesse à l'exécution. Compte tenu de ces constatations, nous proposons dans le cadre de cette thèse une méthode d'optimisation hybride qui vise deux objectifs : la production de plans d'exécution robustes, notamment lorsque l'incertitude des estimations utilisées est importante, et la correction d'une violation de la robustesse pendant l'exécution. Notre méthode s'appuie sur des intervalles d'estimations calculés autour des paramètres incertains, pour produire des plans d'exécution robustes. Ces plans sont ensuite enrichis par des opérateurs dits de contrôle et de décision. Ces opérateurs collectent des statistiques à l'exécution et vérifient la robustesse du plan en cours. Si la robustesse est violée, ces opérateurs sont capables de prendre des décisions de corrections du reste du plan sans avoir besoin de rappeler l'optimiseur. Les résultats de l'évaluation des performances de notre méthode indiquent qu'elle fournit des améliorations significatives dans la robustesse d'évaluation de requêtes. / The quality of an execution plan generated by a query optimizer is highly dependent on the quality of the estimates produced by the cost model. Unfortunately, these estimates are often imprecise. A body of work has been done to improve estimate accuracy. However, obtaining accurate estimates remains very challenging since it requires a prior and detailed knowledge of the data properties and run-time characteristics. Motivated by this issue, two main optimization approaches have been proposed. A first approach relies on single-point estimates to choose an optimal execution plan. At run-time, statistics are collected and compared with estimates. If an estimation error is detected, a re-optimization is triggered for the rest of the plan. At each invocation, the optimizer uses specific values for parameters required for cost calculations. Thus, this approach can induce several plan re-optimizations, resulting in poor performance. In order to avoid this, a second approach considers the possibility of estimation errors at the optimization time. This is modelled by the use of multi-point estimates for each error-prone parameter. The aim is to anticipate the reaction to a possible plan sub-optimality. Methods in this approach seek to generate robust plans, which are able to provide good performance for several run-time conditions. These methods often assume that it is possible to find a robust plan for all expected run-time conditions. This assumption remains unjustified. Moreover, the majority of these methods maintain without modifications an execution plan until the termination. This can lead to poor performance in case of robustness violation at run-time. Based on these findings, we propose in this thesis a hybrid optimization method that aims at two objectives : the production of robust execution plans, particularly when the uncertainty in the used estimates is high, and the correction of a robustness violation during execution. This method makes use of intervals of estimates around error-prone parameters. It produces execution plans that are likely to perform reasonably well over different run-time conditions, so called robust plans. Robust plans are then augmented with what we call check-decide operators. These operators collect statistics at run-time and check the robustness of the current plan. If the robustness is violated, check-decide operators are able to make decisions for plan modifications to correct the robustness violation without a need to recall the optimizer. The results of performance studies of our method indicate that it provides significant improvements in the robustness of query processing.

Le procédé de fabrication de mélanges (PM) consiste à déterminer les proportions optimales à mélanger d'un ensemble de composants de façon que le produit obtenu satisfasse un ensemble de spécifications sur leurs propriétés. Deux caractéristiques importantes du problème de mélange sont les bornes dures sur les propriétés du mélange et l'incertitude répandue dans le procédé. Dans ce travail, on propose une méthode pour la production de mélanges robustes en temps réel qui minimise le coût de la recette et la sur-qualité du mélange. La méthode est basée sur les techniques de l'Optimisation Robuste et sur l'hypothèse que les lois des mélange sont linéaires. On exploite les polytopes sous-jacents pour mesurer, visualiser et caractériser l'infaisabilité du PM et on analyse la modification des bornes sur les composants pour guider le procédé vers le ``meilleur`` mélange robuste. On propose un ensemble d'indicateurs et de visualisations en vue d'offrir une aide à la décision. / The oil blending process (BP) consists in determining the optimal proportions to blend from a set of available components such that the final product fulfills a set of specifications on their properties. Two important characteristics of the blending problem are the hard bounds on the blend's properties and the uncertainty pervading the process. In this work, a real-time optimization method is proposed for producing robust blends while minimizing the blend quality giveaway and the recipe's cost. The method is based on the Robust Optimization techniques and under the assumption that the components properties blend linearly. The blending intrinsic polytopes are exploited in order to measure, visualize and characterize the infeasibility of the BP. A fine analysis of the components bounds modifications is conducted to guide the process towards the ``best`` robust blend. A set of indices and visualizations provide a helpful support for the decision maker.

Dans un marché électrique concurrentiel, EDF a adapté ses outils de gestion de production pour permettre une gestion optimale de son portefeuille, particulièrement sur les horizons journaliers et infra-journaliers, derniers leviers pour une gestion optimisée de la production. Et plus l'horizon d'optimisation s'approche du temps réel, plus les décisions prises aux instants précédents deviennent structurantes voire limitantes en terme d'actions. Ces décisions sont aujourd'hui prises sans tenir compte du caractère aléatoire de certaines entrées du modèle. En effet, pour les décisions à court-terme, la finesse et la complexité des modèles déjà dans le cas déterministe ont souvent été un frein à des travaux sur des modèles tenant compte de l'incertitude. Pour se prémunir face à ces aléas, des techniques d'optimisation en contexte incertain ont fait l'objet des travaux de cette thèse. Nous avons ainsi proposé un modèle robuste de placement de la production tenant compte des incertitudes sur la demande en puissance. Nous avons construit pour cette fin un ensemble d'incertitude permettant une description fine de l'aléa sur les prévisions de demande en puissance. Le choix d'indicateurs fonctionnels et statistiques a permis d'écrire cet ensemble comme un polyèdre d'incertitude. L'approche robuste prend en compte la notion de coût d'ajustement face à l'aléa. Le modèle a pour objectif de minimiser les coûts de production et les pires coûts induits par l'incertitude. Ces coûts d'ajustement peuvent décrire différents contextes opérationnels. Une application du modèle robuste à deux contextes métier est menée avec un calcul du coût d'ajustement approprié à chaque contexte. Enfin, le présent travail de recherche se situe, à notre connaissance, comme l'un des premiers dans le domaine de la gestion optimisée de la production électrique à court terme avec prise en compte de l'incertitude. Les résultats sont par ailleurs susceptibles d'ouvrir la voie vers de nouvelles approches du problème. / Robust Optimization is an approach typically offered as a counterpoint to Stochastic Programming to deal with uncertainty, especially because it doesn't require any precise information on stochastic distributions of data. In the present work, we deal with challenging unit-commitment problem for the French daily electricity production under demand uncertainty. Our contributions concern both uncertainty modelling and original robust formulation of unit-commitment problem. We worked on a polyhedral set to describe demand uncertainty, using statistical tools and operational indicators. In terms of modelling, we proposed robust solutions that minimize production and worst adjustment costs due to uncertainty observation. We study robust solutions under two different operational contexts. Encouraging results to the convex unit-commitment problems under uncertainty are thus obtained, with intersting research topics for future work.

A ascensão de um grande número de pessoas em países em desenvolvimento para a classe média, no inicio do século XXI, aliado ao movimento político para transferência de base energética para os biocombustíveis vêm aumentando a pressão sobre os preços das commodities agrícolas e apresentando novas oportunidades e cenários administrativos para os produtores agrícolas dessas commodities, em especial aquelas que podem se diversificar em muitos subprodutos para atender diferentes mercados, como o de alimentos, químico, têxtil e de energia. Nesse novo ambiente os produtores podem se beneficiar dividindo adequadamente a produção entre os diferentes subprodutos, definindo o melhor momento para a comercialização através de estoques, e ainda controlar sua exposição ao risco através de posições no mercado de derivativos. A literatura atual pouco aborda o tema da diversificação e seu impacto nas decisões de produção e comercialização agrícola e portanto essa tese tem o objetivo de propor um modelo de decisão fundado na teoria de seleção de portfólios capaz de decidir a divisão da produção entre diversos subprodutos, as proporções a serem estocadas e o momento mais adequado para a comercialização e por fim as posições em contratos futuros para fins de proteção ou hedge. Adicionalmente essa tese busca propor que esse modelo seja capaz de lidar com incerteza em parâmetros, em especial parâmetros que provocam alto impacto nos resultados, como é o caso dos retornos previstos no futuro. Como uma terceira contribuição, esse trabalho busca ainda propor um modelo de previsão de preços mais sofisticado que possa ser aplicado a commodities agrícolas, em especial um modelo híbrido ou hierárquico, composto de dois modelos, um primeiro modelo fundado sob a teoria de processos estocásticos e do Filtro de Kalman e um segundo modelo, para refinar os resultados do primeiro modelo de previsão, baseado na teoria de redes neurais, com a finalidade de considerar variáveis exógenas. O modelo híbrido de previsão de preços foi testado com dados reais do mercado sucroalcooleiro brasileiro e indiano, gerando resultados promissores, enquanto o modelo de decisão de parâmetros de produção, comercialização, estocagem e hedge se mostrou uma ferramenta útil para suporte a decisão após ser testado com dados reais do mercado sucroalcooleiro brasileiro e do mercado de milho, etanol e biodiesel norte-americano. / The rise of a large number of people in developing countries for the middle class at the beginning of the century, combined with the political movement to transfer the energy base for biofuels has been increasing pressure on prices of agricultural commodities and presenting new opportunities and administrative scenarios for agricultural producers of these commodities, especially those who may diversify into many products to meet different markets such as food, chemicals, textiles and energy. In this new environment producers can achieve benefits properly dividing production between different products, setting the best time to market through inventories, and still control their risk exposure through positions in the derivatives market. The literature poorly addresses the issue of diversification and its impact on agricultural production and commercialization decisions and therefore this thesis aims to propose a decision model based on the theory of portfolio selection able to decide the division of production between different products, the proportions to be stored and timing for marketing and finally the positions in futures contracts to hedge. Additionally this thesis attempts to propose that this model is capable of dealing with uncertainty in parameters, especially parameters that cause high impact on the results, as is the case of expected returns in the future. As a third contribution this paper seeks to also propose a model more sophisticated to forecast prices that can be applied to agricultural commodities, especially a hybrid or hierarchical model, composed of two models, a first one based on the theory of stochastic processes and Kalman filter and a second one to refine the results of the first prediction model, based on the theory of neural networks in order to consider the exogenous variables. The hybrid model for forecasting prices has been tested with real data from the Brazilian and Indian sugar ethanol market, generating promising results, while the decision model parameters of production, commercialization, storage and hedge proved a useful tool for decision support after being tested with real data from Brazilian sugar ethanol market and the corn, ethanol and biodiesel market in U.S.A.

Orientador: Alberto Luiz Francato / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Civil, Arquitetura e Urbanismo / Made available in DSpace on 2018-08-23T22:42:20Z (GMT). No. of bitstreams: 1 Deantoni_VictordeBarros_M.pdf: 2817178 bytes, checksum: 1c05a50aa818606a82c052b88f62c14d (MD5) Previous issue date: 2013 / Resumo: O planejamento da operação do setor elétrico brasileiro é realizado com base em modelos que por meio de otimização determinam a geração de energia de fontes térmicas e hidráulicas. Utilizando a técnica de otimização estocástica robusta possibilita-se a análise com um conjunto de cenários históricos, com o objetivo de determinar a operação do primeiro intervalo de tempo do horizonte de planejamento, e assim obter uma solução que não necessariamente seja a melhor para um determinado cenário, mas sim uma solução que seja interessante para qualquer um dos cenários que possam ocorrer. O objetivo deste trabalho foi criar uma nova versão do modelo SolverSIN com um módulo de otimização estocástica robusta, chamado de SolverSINR, esse novo modelo permite o planejamento da operação considerando um conjunto de cenários históricos. As séries históricas são obtidas do relatório do programa mensal de operação, que é um arquivo de saída do NEWAVE. Para organização dessas séries foi desenvolvido um aplicativo chamado SHENA. O novo modelo apresentou resultados coerentes em comparação com o modelo SolverSIN determinístico e mostrou valores viáveis para a operação de reservatórios. A ferramenta permite a otimização assumindo a hipótese de repetição de uma série histórica. O modelo SolverSINR vem a contribuir como mais uma alternativa de avaliação crítica às políticas operacionais do SIN implementadas pelos modelos oficiais do SEB. Durante a avaliação de resultados notou-se que uma restrição de geração hidráulica mínima para o subsistema Nordeste causava infactibilidade em alguns cenários, adotou-se uma variável de folga para solucionar esse problema. Destaca-se que houve factibilidade nos procedimentos de operação em tempo de processamento compatível com otimização estocástica de processos em equipamentos usais para tal tarefa / Abstract: The operation planning of the interconnected power generation Brazilian system is carried out based on models through deterministic optimization power generation from thermal and hydraulic sources. Using the robust stochastic optimization technique enables the analysis grounded in a set of historical scenarios, in order to determine the operation of the first time interval of the planning horizon, and thus obtain a solution that is not necessarily the best for a selected scenario, but a solution that is interesting for any of the scenarios that might happen. The objective of this work was to create a new version of the model SolverSIN with a robust stochastic optimization module, called SolverSINR , this new model allows for the operation planning considering a set of historical scenarios . The time series is obtained from the monthly report called "pmo.dat" that is an output file from NEWAVE model. For organizing these series was developed an application called SHENA. The new model showed consistent results in comparison with the deterministic model (SolverSIN) and showed feasible values for reservoir operation. The tool allows the optimization under the assumption of repetition of a historical series. The model SolverSINR comes to contribute as an alternative assessment to critical operational policies implemented by SIN official models of SEB. During the evaluation of results was noted that a restriction of minimum hydraulic generation on Northeast subsystem caused infeasible answer in some scenarios, we adopted a penalty variable to solve this problem. It is noteworthy that there was feasibility in operating procedures in processing time compatible with stochastic optimization of processes in usual equipment for this task / Mestrado / Recursos Hidricos, Energeticos e Ambientais / Mestre em Engenharia Civil

No description available.