In Bayesian decision analysis, uncertainty and risk are accounted for with probabilities for the possible states, or states of nature, that affect the outcome of a decision. Application of Bayes’ theorem requires non-informative prior probabilities, which represent the probabilities of states of nature for a decision maker under complete ignorance. These prior probabilities are then subsequently updated with any and all available information in assessing probabilities for making decisions. The conventional approach for the non-informative probability distribution is based on Bernoulli’s principle of insufficient reason. This principle assigns a uniform distribution to uncertain states when a decision maker has no information about the states of nature. The principle of insufficient reason has three difficulties: it may inadvertently provide a biased starting point for decision making, it does not provide a consistent set of probabilities, and it violates reasonable axioms of decision theory. The first objective of this study is to propose and describe a new method to establish non-informative prior probabilities for decision making under uncertainty. The proposed decision-based method is focuses on decision outcomes that include preference in decision alternatives and decision consequences. The second objective is to evaluate the logic and rationality basis of the proposed decision-based method. The decision-based method overcomes the three weaknesses associated with the principle of insufficient reason, and provides an unbiased starting point for decision making. It also produces consistent non-informative probabilities. Finally, the decision-based method satisfies axioms of decision theory that characterize the case of no information (or complete ignorance). The third and final objective is to demonstrate the application of the decision-based method to practical decision making problems in engineering. Four major practical implications are illustrated and discussed with these examples. First, the method is practical because it is feasible in decisions with a large number of decision alternatives and states of nature and it is applicable to both continuous and discrete random variables of finite and infinite ranges. Second, the method provides an objective way to establish non-informative prior probabilities that capture a highly nonlinear relationship between states of nature. Third, we can include any available information through Bayes’ theorem by updating the non-informative probabilities without the need to assume more than is actually contained in the information. Lastly, two different decision making problems with the same states of nature may have different non-informative probabilities. / text

La modélisation cardiaque personnalisée consiste à créer des simulations 3D virtuelles de cas cliniques réels pour aider les cliniciens à prédire le comportement du cœur ou à mieux comprendre certaines pathologies. Dans cette thèse nous illustrons d'abord la nécessité d'une approche robuste d'estimation des paramètres, dans un cas ou l'incertitude dans l'orientation des fibres myocardiques entraîne une incertitude dans les paramètres estimés qui est très large par rapport à leur variabilité physiologique. Nous présentons ensuite une approche originale multi-échelle 0D/3D pour réduire le temps de calcul, basée sur un couplage multi-échelle entre les simulations du modèle 3D et d'une version "0D" réduite de ce modèle. Ensuite, nous dérivons un algorithme rapide de personnalisation multi-échelle pour le modèle 3D. Dans un deuxième temps, nous construisons plus de 140 simulations 3D personnalisées, dans le cadre de deux études impliquant l'analyse longitudinale de la fonction cardiaque : d'une part, l'analyse de l'évolution de cardiomyopathies à long terme, d'autre part la modélisation des changements cardiovasculaires pendant la digestion. Enfin, nous présentons un algorithme pour sélectionner automatiquement des directions observables dans l'espace des paramètres à partir d'un ensemble de mesures, et calculer des probabilités "a priori" cohérentes dans ces directions à partir des valeurs de paramètres dans la population. Cela permet en particulier de contraindre l'estimation de paramètres dans les cas où des mesures sont manquantes. Au final nous présentons des estimations cohérentes de paramètres dans une base de données de 811 cas avec le modèle 0D et 137 cas du modèle 3D. / Personalised cardiac modeling consists in creating virtual 3D simulations of real clinical cases to help clinicians predict the behaviour of the heart, or better understand some pathologies from the estimated values of biophysical parameters. In this work we first motivate the need for a consistent parameter estimation framework, from a case study were uncertainty in myocardial fibre orientation leads to an uncertainty in estimated parameters which is extremely large compared to their physiological variability. To build a consistent approach to parameter estimation, we then tackle the computational complexity of 3D models. We introduce an original multiscale 0D/3D approach for cardiac models, based on a multiscale coupling to approximate outputs of a 3D model with a reduced "0D" version of the same model. Then we derive from this coupling an efficient multifidelity optimisation algorithm for the 3D model. In a second step, we build more than 140 personalised 3D simulations, in the context of two studies involving the longitudinal analysis of the cardiac function: on one hand the analysis of long-term evolution of cardiomyopathies under therapy, on the other hand the modeling of short-term cardiovascular changes during digestion. Finally we present an algorithm to automatically detect and select observable directions in the parameter space from a set of measurements, and compute consistent population-based priors probabilities in these directions, which can be used to constrain parameter estimation for cases where measurements are missing. This enables consistent parameter estimations in a large databases of 811 cases with the 0D model, and 137 cases of the 3D model.

This paper concerns about key business process in the production companies, namely, the new product development. The object of this thesis has been to create a tool to estimate the risk of the new product development. To reach this goal, current tools used to deciding the risk must have been explored. As the best tool, appropriate for assessing the risk of new product development has proved the Bayesian Network. This paper explains the construction of the Bayesian network and shows the way how to generate the probabilities in the network to be accurate for the risk estimation. Based on this theoretical knowledge has been built an information system, which estimates the risk of the new products and administer the risks.