In clustered survival data, unobservable cluster effects may exert powerful influences on the outcomes and thus induce correlation among subjects within the same cluster. The ordinary partial likelihood approach does not account for this dependence. Frailty models, as an extension to Cox regression, incorporate multiplicative random effects, called frailties, into the hazard model and have become a very popular way to account for the dependence within clusters. We particularly study the two-level nested lognormal frailty model and propose an estimation approach based on the complete data likelihood with frailty terms integrated out. We adopt B-splines to model the baseline hazards and adaptive Gauss-Hermite quadrature to approximate the integrals efficiently. Furthermore, in finding the maximum likelihood estimators, instead of the Newton-Raphson iterative algorithm, Gauss-Seidel and BFGS methods are used to improve the stability and efficiency of the estimation procedure. We also study competing risks models with missing cause of failure in the context of Cox proportional hazards models. For competing risks data, there exists more than one cause of failure and each observed failure is exclusively linked to one cause. Conceptually, the causes are interpreted as competing risks before the failure is observed. Competing risks models are constructed based on the proportional hazards model specified for each cause of failure respectively, which can be estimated using partial likelihood approach. However, the ordinary partial likelihood is not applicable when the cause of failure could be missing for some reason. We propose a weighted partial likelihood approach based on complete-case data, where weights are computed as the inverse of selection probability and the selection probability is estimated by a logistic regression model. The asymptotic properties of the regression coefficient estimators are investigated by applying counting process and martingale theory. We further develop a double robust approach based on the full data to improve the efficiency as well as the robustness.

Ce travail se décompose en trois parties distinctes. Dans la première partie, on introduit un algorithme du type Gauss-Seidel pour la minimisation de fonctionnelles symétriques semi-définies positives. La convergence de cet algorithme est démontrée. En application, on donne deux méthodes de lissage de surfaces. Ces méthodes sont basées sur l'idée de ramener un probleme de lissage a deux dimensions a la resolution d'une suite de problèmes a une dimension faciles a résoudre. Pour cela on utilise l'opération d'inf-convolution spline. Dans la deuxième partie, on introduit une nouvelle methode pour la conception d'un verre progressif. Ce verre est représente par une surface suffisamment régulière, a laquelle on impose des conditions sur ses courbures principales dans certaines zones (zone de vision de loin et zone de vision de pres), et des conditions sur ses directions principales de courbure dans d'autres zones (zone nasale et zone temporale). La surface est écrite sous forme de produit tensoriel de b-splines de degré quatre. Pour la calculer, on est amené a minimiser un opérateur non quadratique. Cette minimisation est alors effectuée par un procédé itératif dont on a teste numériquement la convergence rapide

.

.

Dans la théorie des D-modules, on définit les systèmes de Gauss-Manin par l'image directe par un morphisme du faisceau structural. Un résultat essentiel est leur régularité. Dans cette thèse, on s'intéresse à l'irrégularité d'un analogue des systèmes de Gauss-Manin, l'image directe par un polynôme f d'un D-module élémentaire associé à un polynôme g, essentiellement dans le cas à deux variables. On utilisera deux approches que l'on comparera. Cette irrégularité permet de contrôler la croissance non modérée des intégrales d'une forme algébrique relative sur une collection de classes d'homologie dans les fibres de f, localement constante à supports fermés convenablement choisis. Dans une première méthode, nous exprimerons l'irrégularité en c de ces systèmes à l'aide de la courbe discriminante de f et g. On utilisera pour cela les travaux de Lê Dung Trang et C. Weber sur les résolutions à l'infini. En utilisant le théorème de commutation dû à Z. Mebkhout de l'image directe avec le complexe d'irrégularité, on se ramène alors aux calculs de caractéristiques d'Euler de complexes d'irrégularité de D-modules à deux variables dont le lieu singulier est un croisement normal. Un résultat de C. Sabbah permet alors de lier ces caractéristiques d'Euler à celles d'une fibre de Milnor. Pour l'irrégularité à l'infini, il faut ajouter une courbe spéciale qui provient des diviseurs dicritiques pour f et g d'une résolution à l'infini. Dans une deuxième méthode, on se ramènera au cas où f et g sont des projections. On exprimera alors l'irrégularité en fonction des cycles caractéristiques du complexe image directe par (f,g) du faisceau structural.

We are examining iteratively regularized methods for solving nonlinear inverse problems. Of particular interest for these types of methods are application problems which are unstable. For these application problems, special methods of numerical analysis are necessary, since classical algorithms tend to be divergent.

This paper presents a study of the computer game ”Minesweeper”. The aim of the game is to search through a rectangular area of mined squares without hitting any mines. By using a strategy based on making every operation as safe as possible, series of the game have been simulated. The size of the playground is four times four squares. The si- mulations indicate how often the game will succeed and which square is the best one to start at. The strategy demands advanced mathematical calculations. The account of these is the ma- jor part of my work. My investigation shows that if there are three hidden mines the game will succeed about two times out of three. It also shows that the best startingpoint is a corner.

In this paper, we study the parameter estimation of the multiple linear time series regression model with long memory stochastic regressors and innovations. Robinson and Hidalgo (1997) and Hidalgo and Robinson (2002) proposed a class of frequency-domain weighted least squares estimates. Their estimates are shown to achieve the Gauss-Markov bound with standard convergence rate. In this study, we proposed a time-domain generalized LSE approach, in which the inverse autocovariance matrix of the innovations is estimated via autoregressive coefficients. Simulation studies are performed to compare the proposed estimates with Robinson and Hidalgo (1997) and Hidalgo and Robinson (2002). The results show the time-domain generalized LSE is comparable to Robinson and Hidalgo (1997) and Hidalgo and Robinson (2002) and attains higher efficiencies when the autoregressive or moving average coefficients of the FARIMA models have larger values. A variance reduction estimator, called TF estimator, based on linear combination of the proposed estimator and Hidalgo and Robinson (2002)'s estimator is further proposed to improve the efficiency. Bootstrap method is applied to estimate the weights of the linear combination. Simulation results show the TF estimator outperforms the frequency-domain as well as the time-domain approaches.

The field of Computational Fluid Dynamics (CFD) is in a continual state of advancement due to new numerical techniques, optimization of existing codes, and the increase in memory and processing speeds of computers. In this thesis, the solution technique for a pre-existing Navier-Stokes flow solver is adapted from an explicit Runge Kutta method to a Lower-Upper Symmetric Gauss-Seidel (LU-SGS) implicit time integration method. Explicit time integration methods were originally used in CFD codes because these methods require less memory. Information needed to advance the flow in time is localized to each grid point. These explicit methods are, however, restricted by time step sizes due to stability criteria. In contrast, implicit methods are unaffected by a large time step sizes but are restricted by memory requirements due to the complexities of unstructured grids. The implementation of LU-SGS performs grid re-ordering for unstructured meshes because of the coupling of grid points in the integration method's solution. The explicit and implicit flow solvers were tested for inviscid flows in incompressible, compressible, and transoinc flow regimes. The results found by comparing the implicit and explicit algorithms revealed a significant speed up in convergence to steady state by the LU-SGS method in terms of iteration number and CPU time per iteration.

In this thesis, we have a comprehensive study of the famous compound Gauss-Markov image model. In this model, a pixel in the image random field is determined by the surrounding pixels according to a predetermined line field. This model is useful in image restoration by applying two steps iteratively: restoring the line field by the assumed image field and restoring the image field by the just computed line field. CGM (Compound Gauss-Markov) image modeling is characterized by the line fields and the generating noise. In this thesis we apply combinations of techniques such as changing processing order, immediate updating, probability determination and different methods to find the best modeling. Furthermore, the effects of the above modeling are demonstrated by its energy, visual quality, and error resistance. Finally, by solving a set of nonlinear equations we apply the CGM model to an image restoration problem for image corrupted by a dusted lens.