[pt] Face à complexidade da matriz energética brasileira e em particular de todo o sistema elétrico de potência interligado, torna-se imprescindível garantir que os equipamentos instalados, desde a geração até os consumidores finais, operem em condições satisfatórias e em elevados níveis de confiabilidade. De acordo com a reestruturação do setor e sua inserção em um mercado competitivo, além da elevada confiabilidade exigida, o novo conceito de disponibilidade dos equipamentos e sistemas impõe mais qualidade e planejamento no mercado de energia elétrica. Neste contexto, as empresas de energia elétrica preocupam-se cada vez mais em manter seus equipamentos em boas condições de operação para que tais metas sejam alcançadas. Entre os equipamentos elétricos de potência, os transformadores sem dúvida correspondem ao ativo de maior importância por serem os mais caros e complexos em termos funcionais. A Análise de Gases Dissolvidos no óleo mineral isolante (AGD) é uma ferramenta de diagnóstico de grande aceitação e potencial na detecção de faltas em equipamentos elétricos com isolação papel-óleo, sobretudo nos transformadores de potência. Com o objetivo de fornecer maiores subsídios aos gestores de manutenção na tomada de decisões quanto à intervenções em transformadores, o trabalho desenvolvido propõe um Sistema de Apoio à Decisão composto por um módulo de Inteligência Computacional (IC) que utiliza regras fuzzy para efetuar o diagnóstico do equipamento em conjunto com outro módulo de apoio à decisão que considera as características do transformador e outros parâmetros de influência, fornecendo, além do diagnóstico, recomendações para a tomada de decisão pelos gestores de manutenção. / [en] Given the complexity of the Brazilian energy matrix and in particular of the whole interconnected electrical power system, it is essential to ensure that the equipment, from generation to final consumers, operates in a satisfactory way and high levels of reliability. In accordance with the restructuring of the sector and its integration in a competitive market, in addition to the high reliability required, the new concept of availability of equipment and systems requires more planning and quality in the market of electrical energy. In this context, electrical power companies are increasingly concerned about maintaining their equipment in good operating conditions so that the above targets are attained. Within electrical equipments, power transformers are undoubtedly the most important asset, since they are the most expensive and complex in functional terms. Dissolved Gases Analysis in insulating mineral oil (DGA), is a widely accepted tool for detecting faults in electrical equipments with paper-oil insulation, particularly in power transformers. With the aim of providing more subsidies to maintenance managers when making decisions on interventions in power transformers, this work proposes a Decision Support System composed of a module of Computational Intelligence (CI) which uses fuzzy rules to diagnose the equipment, together with another decision-support module, which considers the power transformer features and other parameters in order to help managers in the decision making process.

Ce travail de thèse s'inscrit autour des enjeux visant à évaluer les impacts des décisions d'aménagement à l'aune des impératifs de développement durable. Au delà du flou se dégageant souvent de ce concept, il s'agit ici de conduire une réflexion sur un meilleur équilibre à trouver entre les trois sphères généralement identifiées comme piliers du développement durable : l'économique, le social et l'environnemental. L’objectif s’inscrit dans la réflexion suivante : comment évaluer les impacts des politiques d’aménagement dans chacune de ces trois sphères, et quelles conclusions en tirer ?Il s'agit ici d'apporter des éléments de réponse en évaluant la performance des territoires étudiés dans ces trois sphères du développement durable. Pour cela, le travail s'appuie sur les sorties des modèles de simulation (ici, le modèle MobiSim développé au sein du laboratoire ThéMA) et la production d'indicateurs synthétiques permettant une analyse et une évaluation lisible de l'espace. La combinaison de ces indicateurs synthétiques permet de visualiser et d'analyser la durabilité du territoire étudié et d'en déduire les mesures appropriées à mettre en œuvre en vue d'assurer une politique de développement durable.Il s'agit ainsi de mettre en perspective la méthodologie et les résultats obtenus dans une vision globale, cherchant à établir par l'équilibre entre les trois sphères une harmonie qui permette de satisfaire aux objectifs d'une politique guidée par les impératifs liés au concept de développement durable hérités du rapport Bruntland. Une approche critique de ce concept et une analyse des jeux d'acteurs d'un territoire sont ici conduits de manière à pouvoir implémenter les travaux de recherche scientifique au sein d'un processus de décision. L'enjeu est ainsi de permettre aux résultats des travaux de recherche basés sur les outils de modélisation de pouvoir trouver une issue plus concrète et plus opérationnelle. / This thesis work is based on issues aimedat assessing the impacts of planning decisions in the light of the imperatives of sustainable development. Beyond the vagueness that often emerges from this concept, the aim here is to reflect on a better balance to be found between the three fields generally identified as pillars of sustainable development : the economic, the social and the environmental. The objective is part of the following reflection: how to evaluate the impacts of management policies in each of these three pillars, and what conclusions to draw from them ? The aim here is to provide answers by assessing the performance of the territories studied in these three pillars of sustainable development. For this, the work relies on the outputs of the simulation models (here, the MobiSim model developed within the ThéMA laboratory) and the production of synthetic indicators allowing analysis and a readable evaluation of the space.The combination of these synthetic indicators makes it possible to visualize and analyze the sustainability of the territory studied and to deduce the appropriate measures to implement in order to ensure asustainable development policy. It is thus a question of putting in perspective the methodology and the results obtained in a global vision, seeking to establish by the balance between the three spheres a harmony which makes it possible to satisfy the objectives of a policy guided by the imperatives related to the sustainable development concept inherited from the Brundtland report. A critical approach to this concept and an analysis of the games of actors of a territory are conducted here in order to implement the scientific research work within a decision process. The challenge is to enable results of research work based on modeling tools to be able to find a more concrete and more operational outcom.

This dissertation discusses three problems from different areas of medical research and their machine learning solutions. Each solution is a distinct type of decision support system. They show three common properties: personalized healthcare decision support, reduction of the use of medical resources, and improvement of outcomes. The first decision support system assists individual hospital selection. This system can help a user make the best decision in terms of the combination of mortality, complication, and travel distance. Both machine learning and optimization techniques are utilized in this type of decision support system. Machine learning methods, such as Support Vector Machines, learn a decision function. Next, the function is transformed into an objective function and then optimization methods are used to find the values of decision variables to reach the desired outcome with the most confidence. The second decision support system assists diagnostic decisions in a sequential decision-making setting by finding the most promising tests and suggesting a diagnosis. The system can speed up the diagnostic process, reduce overuse of medical tests, save costs, and improve the accuracy of diagnosis. In this study, the system finds the test most likely to confirm a diagnosis based on the pre-test probability computed from the patient's information including symptoms and the results of previous tests. If the patient's disease post-test probability is higher than the treatment threshold, a diagnostic decision will be made, and vice versa. Otherwise, the patient needs more tests to help make a decision. The system will then recommend the next optimal test and repeat the same process. The third decision support system recommends the best lifestyle changes for an individual to lower the risk of cardiovascular disease (CVD). As in the hospital recommendation system, machine learning and optimization are combined to capture the relationship between lifestyle and CVD, and then generate recommendations based on individual factors including preference and physical condition. The results demonstrate several recommendation strategies: a whole plan of lifestyle changes, a package of n lifestyle changes, and the compensatory plan (the plan that compensates for unwanted lifestyle changes or real-world limitations).

A key factor of software architecting is the decision-making process. All phases of software development contain some kind of decision-making activities. However, the software architecture decision process is the most challenging part. To support the decision-making process, a research project named ORION provided a knowledge repository that contains a collection of decision cases. To utilize the collected data in an efficient way, eliciting correlations between decision cases needs to be automated.  The objective of this thesis is to select appropriate method(s) for automatically detecting correlations between decision cases. To do this, an experiment was conducted using a dataset of collected decision cases that are based on a taxonomy called GRADE. The dataset is stored in the Neo4j graph database. The Neo4j platform provides a library of graph algorithms which allow to analyse a number of relationships between connected data. In this experiment, five Similarity algorithms are used to find correlated decisions, then the algorithms are analysed to determine whether the they would help improve decision-making.  From the results, it was concluded that three of the algorithms can be used as a source of support for decision-making processes, while the other two need further analyses to determine if they provide any support.

Drug effects can be unpredictable and vary widely among patients with environmental, genetic, and clinical factors. Randomized control trials (RCTs) are not sufficient to identify adverse drug reactions (ADRs), and the electronic health record (EHR) along with medical claims have become an important resource for pharmacovigilance. Among all the data collected in hospitals, laboratory tests represent the most documented and reliable data type in the EHR. Laboratory tests are at the core of the clinical decision process and are used for diagnosis, monitoring, screening, and research by physicians. They can be linked to drug effects either directly, with therapeutic drug monitoring (TDM), or indirectly using drug laboratory effects (DLEs) that affect surrogate tests. Unfortunately, very few automated methods use laboratory tests to inform clinical decision making and predict drug effects, partly due to the complexity of these time series that are irregularly sampled, highly dependent on other clinical covariates, and non-stationary. Deep learning, the branch of machine learning that relies on high-capacity artificial neural networks, has known a renewed popularity this past decade and has transformed fields such as computer vision and natural language processing. Deep learning holds the promise of better performances compared to established machine learning models, although with the necessity for larger training datasets due to their higher degrees of freedom. These models are more flexible with multi-modal inputs and can make sense of large amounts of features without extensive engineering. Both qualities make deep learning models ideal candidate for complex, multi-modal, noisy healthcare datasets. With the development of novel deep learning methods such as generative adversarial networks (GANs), there is an unprecedented opportunity to learn how to augment existing clinical dataset with realistic synthetic data and increase predictive performances. Moreover, GANs have the potential to simulate effects of individual covariates such as drug exposures by leveraging the properties of implicit generative models. In this dissertation, I present a body of work that aims at paving the way for next generation laboratory test-based clinical decision support systems powered by deep learning. To this end, I organized my experiments around three building blocks: (1) the evaluation of various deep learning architectures with laboratory test time series and their covariates with a forecasting task; (2) the development of implicit generative models of laboratory test time series using the Wasserstein GAN framework; (3) the inference properties of these models for the simulation of drug effects in laboratory test time series, and their application for data augmentation. Each component has its own evaluation: The forecasting task enabled me to explore the properties and performances of different learning architectures; the Wasserstein GAN models are evaluated with both intrinsic metrics and extrinsic tasks, and I always set baselines to avoid providing results in a "neural-network only" referential. Applied machine learning, and more so with deep learning, is an empirical science. While the datasets used in this dissertation are not publicly available due to patient privacy regulation, I described pre-processing steps, hyper-parameters selection and training processes with reproducibility and transparency in mind. In the specific context of these studies involving laboratory test time series and their clinical covariates, I found that for supervised tasks, machine learning holds up well against deep learning methods. Complex recurrent architectures like long short-term memory (LSTM) do not perform well on these short time series, while convolutional neural networks (CNNs) and multi-layer perceptrons (MLPs) provide the best performances, at the cost of extensive hyper-parameter tuning. Generative adversarial networks, enabled by deep learning models, were able to generate high-fidelity laboratory test time series, and the quality of the generated samples was increased with conditional models using drug exposures as auxiliary information. Interestingly, forecasting models trained on synthetic data exclusively still retain good performances, confirming the potential of GANs in privacy-oriented applications. Finally, conditional GANs demonstrated an ability to interpolate samples from drug exposure combinations not seen during training, opening the way for laboratory test simulation with larger auxiliary information spaces. In specific cases, augmenting real training sets with synthetic data improved performances in the forecasting tasks, and could be extended to other applications where rare cases present a high prediction error.

Abstract not available

Abstract not available

Abstract not available

In recent years, the popularity of wind energy has grown. It is starting to play a large role in generating renewable, clean energy around the world. In New Zealand, there is increasing recognition and awareness of global warming and the pollution caused by burning fossil fuels, as well as the increased difficulty of obtaining oil from foreign sources, and the fluctuating price of non-renewable energy products. This makes wind energy a very attractive alternative to keep New Zealand clean and green. There are many issues involved in wind farm development. These issues can be grouped into two categories - economic issues and environmental issues. Wind farm developers often use site selection process to minimise the impact of these issues. This thesis aims to develop GIS based models that provide effective decision support tool for evaluating, at a regional scale, potential wind farm locations. This thesis firstly identifies common issues involved in wind farm development. Then, by reviewing previous research on wind farm site selection, methods and models used by academic and corporate sector to solve issues are listed. Criteria for an effective decision support tool are also discussed. In this case, an effective decision support tool needs to be flexible, easy to implement and easy to use. More specifically, an effective decision support tool needs to provide users the ability to identify areas that are suitable for wind farm development based on different criteria. Having established the structure and criteria for a wind farm analysis model, a GIS based tool was implemented using AML code using a Boolean logic model approach. This method uses binary maps for the final analysis. There are a total of 3645 output maps produced based on different combination of criteria. These maps can be used to conduct sensitivity analysis. This research concludes that an effective GIS analysis tool can be developed for provide effective decision support for evaluating wind farm sites.

Existing patient medical records are a rich data source with a potential to support clinical research. Fragmentation of data across disparate medical database inhibits the use of these existing datasets. Overcoming such disjointedness is possible through the use of a data warehouse. Once the data is cleansed, transformed, and stored within the data warehouse it is possible to turn attention to the exploration of the medical datasets. Exploratory and confirmatory Data Mining Tools are well suited to such activities. This thesis concerned with: demonstrating parallels between scientific method and CRISP-DM; extending CRISP-DM for use with medical datasets; and proposal of the supporting Intelligent Decision Support System framework. This research has been undertaken using a fetal-maternal case study. / Master of Science (Hons)