vii, 200 leaves : ill ; 30 cm. / Title page, contents and abstract only. The complete thesis in print form is available from the University Library. / Thesis (Ph.D.)--University of Adelaide, 1989

The time delay estimation between time series is a real-world problem in gravitational lensing, an area of astrophysics. Lensing is the most direct method of measuring the distribution of matter, which is often dark, and the accurate measurement of time delays set the scale to measure distances over cosmological scales. For our purposes, this means that we have to estimate a time delay between two or more noisy and irregularly sampled time series. Estimations have been made using statistical methods in the astrophysics literature, such as interpolation, dispersion analysis, discrete correlation function, Gaussian processes and Bayesian method, among others. Instead, this thesis proposes a kernel-based approach to estimating the time delay, which is inspired by kernel methods in the context of statistical and machine learning. Moreover, our methodology is evolved to perform model selection, regularisation and time delay estimation globally and simultaneously. Experimental results show that this approach is one of the most accurate methods for gaps (missing data) and distinct noise levels. Results on artificial and real data are shown.

Le modèle standard de la cosmologie est fondé sur les hypothèses d'homogénéité et d'isotropie de l'Univers. Pour interpréter la plupart des observations, ces deux hypothèses sont appliquées de façon stricte ; l’objectif principal de cette thèse a été d'évaluer leur pertinence, en particulier lorsque de très petites échelles sont mises en jeu. Après une revue détaillée des lois de l'optique géométrique en espace-temps courbe, on propose une analyse exhaustive de la propagation de la lumière à travers des modèles cosmologiques « en gruyère », modélisant le caractère grumeleux de l'Univers. L'impact sur l'interprétation du diagramme de Hubble s'avère être faible, en particulier grâce à la constante cosmologique. Lorsqu'appliquées aux données actuelles issues de l'observation de supernovae, les corrections associées tendent à améliorer l'accord entre les paramètre cosmologiques mesurés à partir du diagramme de Hubble et du fond diffus cosmologique. Ceci suggère que la précision des observations cosmologiques atteinte aujourd'hui ne permet plus de négliger l'effet des petites structures sur la propagation de lumière à travers le cosmos. Un tel constat a motivé le développement d'un nouveau cadre théorique, inspiré de la physique statistique, visant à décrire ces effets avec précision. Quant à l'hypothèse d'isotropie, cette thèse aborde d'une part les conséquences potentielles d'une anisotropie cosmique sur la propagation de la lumière, en résolvant les équation de l'optique géométrique dans l'espace-temps de Bianchi I. D'autre part, on y analyse une classe de sources d'anisotropie, à savoir les modèles scalaire-vecteur. / The standard model of cosmology is based on the hypothesis that the Universe is spatially homogeneous and isotropic. When interpreting most observations, this cosmological principle is applied stricto sensu ; the main goal of the present thesis was to evaluate how reliable this assumption is, especially when small scales are at stake. After having reviewed the laws of geometric optics in curved spacetime, and the standard interpretation of cosmological observables, the dissertation reports a comprehensive analysis of light propagation in Swiss-cheese models, designed to capture the clumpy character of the Universe. The resulting impact on the interpretation of the Hubble diagram is quantified, and shown to be relatively small, thanks to the cosmological constant. When applied to current supernova data, the associated corrections tend to improve the agreement between the cosmological parameters inferred from the Hubble diagram and from the cosmic microwave background. This is a hint that the effect of small-scale structures on light propagation may become non-negligible in the era of precision cosmology. This motivated the development of a new theoretical framework, based on stochastic processes, which aims at describing small-scale lensing with a better accuracy. Regarding the isotropy side of the cosmological principle, this dissertation addresses, on the one hand, the potential effect of a large-scale anisotropy on light propagation, by solving the equations of geometric optics in the Bianchi I spacetime. On the other hand, possible sources of such an anisotropy, namely scalar-vector models for inflation or dark energy, are analysed.