MSc., Faculty of Science, University of the Witwatersrand, 2011 / The study investigates models and processes that could be used by decision-makers to influence policy. The objectives of the study are to illustrate the considerations in the use of information systems and modeling in the context of sustainable development in South Africa. The problem considered is that of measuring South Africa‟s progress toward sustainable development. The study is conducted at a national level. The design uses statistical indicators arranged into a model based information system using the Bellagio principles of 1996. The statistical indicators are assembled from the United Nations 1993 System of National Accounts, UN 1995 Driver-Pressure-State-Impact-Response framework, United Nations 2003 System of integrated Environmental and Economic Accounting, Basel Committee‟s 2004 Basel II framework, and South African Presidency 2010 Developmental indicators frameworks. The South African economy is conceptualised as a dynamic system composed of five types of capital. The fitted model is a vector autoregressive time series model of order p on a set of statistical factors that describe the South African economy. The robustness of the model to assumptions is evaluated using sensitivity and uncertainty analysis. The optimality of the model output for decision making is evaluated using decision theory. The study will facilitate an evidence based approach to managing South Africa‟s progress towards sustainable development.

Thesis (M.S.) University of Missouri-Columbia, 2006. / The entire dissertation/thesis text is included in the research.pdf file; the official abstract appears in the short.pdf file (which also appears in the research.pdf); a non-technical general description, or public abstract, appears in the public.pdf file. Title from title screen of research.pdf file viewed on (May 20, 2007) Vita. Includes bibliographical references.

Thesis (M.S.)--Ohio University, August, 2004. / Includes bibliographical references (p. 185-187).

Thesis (Ph. D.)--Missouri University of Science and Technology, 2008. / Vita. The entire thesis text is included in file. Title from title screen of thesis/dissertation PDF file (viewed May 28, 2009) Includes bibliographical references.

Thesis (M.S.)--Missouri University of Science and Technology, 2008. / Vita. The entire thesis text is included in file. Title from title screen of thesis/dissertation PDF file (viewed June 5, 2009) Includes bibliographical references (p. 126-132).

Thesis (M.S.)--Ohio University, August, 2004. / Title from PDF t.p. Includes bibliographical references (p. 185-187)

Abstract not available

Thesis (Ph.D.)--McMaster University, 2005. / Supervisor: N.K. Nikolova and M.H. Bakr. Includes bibliographical references (leaves 134-141). Mode of access:World Wide Web.

Global Sensitivity Analysis (GSA) approach helps to identify the effectiveness of model parameters or inputs and thus provides essential information about the model performance. The effects of 14 parameters and one input (forcing data) of the Sacramento Soil Moisture Accounting (SAC-SMA) model are analyzed by using two GSA methods: Sobol' and Fourier Amplitude Sensitivity Test (FAST). The simulations are carried out over five sub-basins within the Columbia River Basin (CRB) for three different periods: one-year, four-year, and seven-year. The main parameter sensitivities (first-order) and the interactions sensitivities (second-order) are evaluated in this study. Our results show that some hydrological processes are highly affected by the simulation length. In other words, some parameters reveal importance during the short period simulation (e.g. one-year) while other parameters are effective in the long period simulations (e.g. four-year and seven-year). Moreover, the reliability of the sensitivity analysis results is compared based on 1) the agreement between the two sensitivity analysis methods (Sobol' and FAST) in terms of highlighting the same parameters or input as the most influential parameters or input and 2) how the methods are cohered in ranking these sensitive parameters under the same conditions (sub-basins and simulation length). The results show that the coherence between the Sobol' and FAST sensitivity analysis methods. Additionally, it is found that FAST method is sufficient to evaluate the main effects of the model parameters and inputs. This study confirms that the Sobol' and FAST methods are reliable GSA methods that can be applied in different scientific applications. Finally, as a future work, we suggest to study the uncertainty associated with the sensitivity analysis approach regarding the reliability of evaluating different sensitivity analysis methods.

La prolifération des électroniques et des émetteurs radiofréquences rend de plus en plus compliqué le processus de conception des systèmes sur le plan CEM. Ce processus doit aboutir à limiter le risque d’interférences ou de défauts au niveau le plus faible notamment dans le contexte des interférences électromagnétiques intentionnelles (IEMI). Ces défauts CEM doivent alors être anticipés lors de la phase de conception. Cependant, du fait de la dispersion des valeurs prises par certains paramètres du système, la modélisation déterministe éprouve quelques difficultés à identifier le risque encouru. La mauvaise connaissance de l’effet des incertitudes associées au système, aboutit alors à prendre des marges de conception considérables conduisant à des surcoûts de fabrication. Pour cette raison, il est devenu important de prendre en compte l’impact des incertitudes des différents paramètres constitutifs d’un système (en phase de conception). Ces paramètres sont essentiellement géométriques (e.g. position de câblages) ou électromagnétiques (e.g. caractéristiques intrinsèques de matériaux). Ils influent par nature sur les performances CEM de ce système. Ces travaux de thèse portent sur l’analyse de la propagation des incertitudes relatives à ces paramètres sur des sorties de modèles de CEM. Le but visé, consiste à quantifier sous une forme probabiliste, le risque de défaut d’un système contenant de nombreux paramètres incertains. Ce type d’étude statistique devrait également permettre, via des analyses de sensibilité, des stratégies de conception de systèmes « fiables » ou à moindres coûts. Dans le contexte des applications visées, les approches dites « fiabilistes » et la méthode dite de « stratification contrôlée », ont été identifiées comme intéressantes, du point de vue de l’analyse d’événements extrêmes. Dans un premier temps, nous nous sommes consacrés à la transposition des méthodes fiabilistes dans un contexte CEM. Ces techniques permettent de quantifier la probabilité de défaillance d’un système, définie comme le dépassement d’un seuil de risque, et renseignent, via une analyse de sensibilité locale, sur les paramètres clés à ajuster. Dans un second temps, nous nous sommes intéressés à la méthode de stratification contrôlée, non appliquée à ce jour à notre connaissance en CEM. L’objectif de cette approche consiste à estimer un quantile extrême de la réponse d’intérêt d’un modèle rigoureux, via l’utilisation d’un modèle simple beaucoup moins coûteux en termes de temps de calcul. Ce processus permet d’accélérer l’obtention d’observations extrêmes, nécessaires à l’estimation du quantile recherché. Les deux techniques ont été mises en oeuvre sur un problème complexe dans un contexte IEMI, pour estimer la probabilité d’occurrence d’événements d’interférences extrêmes. Elles ont permis de dégager des tendances similaires, quant à l’importance de certains paramètres d’entrée incertains sur les événements rares. Les deux méthodes, bien appliquées, pourraient constituer un apport considérable en matière de conception CEM. / The proliferation of electronic and radio frequency transmitters makes more complicated the system design process on the EMC point of view. This process should lead to limit the risk of interferences or defects to lowest level particularly in the context of intentional electromagnetic interferences (IEMI). Therefore, these EMC defects have to be anticipated during the design stage. However, due to the dispersion of the values taken by some parameters of the system, the deterministic modeling presents some difficulties to identify the involved risk. The poor knowledge of the uncertainties effect associated with the system, leads then to take important design margins at the price of additional costs of manufacturing. For this reason, it has become important to take into account the impact of uncertainties of the various constituent parameters of a system (at the design stage). These parameters are essentially geometric (e.g. position of wirings) or electromagnetic (e.g. intrinsic characteristics of materials) ones. They influence by nature the EMC performance of this system. This thesis work deals with the analysis of the propagation of uncertainties of these parameters on EMC model outputs. It aims at quantifying in a probabilistic form, the default risk of a system containing numerous uncertain parameters. This type of statistical analysis should also allow through sensitivity analyses, design strategies of “reliable” systems or at lower cost. In the context of targeted applications, the so-called “reliability approaches” and the “controlled stratification” method have been identified as interesting from the point of view of the analysis of extreme events. Firstly, we are dedicated to the transposition of reliability methods in an EMC context. These techniques are used to quantify the probability of failure of a system, defined as the probability of exceeding a threshold of risk. They inform through a local sensitivity analysis, on the key parameters to adjust. Secondly, we have focused our work on the controlled stratification method, not yet applied in EMC as far as we know. The objective of this approach is to estimate an extreme quantile of the interest response of a rigorous model, using of a much cheaper simple model in terms of computation time. This process allows to speed up the identification of extreme observations required for the estimation of the researched quantile. Both techniques have been applied on a complex problem in an IEMI context, to estimate the probability of occurrence of extreme interference events. They have revealed similar trends as regards to the importance of some uncertain input parameters on rare events. Both methods, properly applied, could provide a significant contribution in terms of EMC design strategy.