Nowadays, although MapReduce is applied to the parallel processing on big data, it has some limitations: for instance, lack of generic but efficient and richly functional primitive parallel methods, incapability of entering multiple input parameters on the entry of parallel methods, and inefficiency in the way of handling iterative algorithms. Spark and Stratosphere are developed to deal with (partly) the shortcoming of MapReduce. The goal of this thesis is to evaluate Spark and Stratosphere both from the point of view of theoretical programming model and practical execution on specified application algorithms. In the introductory section of comparative programming models, we mainly explore and compare the features of Spark and Stratosphere that overcome the limitation of MapReduce. After the comparison in theoretical programming model, we further evaluate their practical performance by running three different classes of applications and assessing usage of computing resources and execution time. It is concluded that Spark has promising features for iterative algorithms in theory but it may not achieve the expected performance improvement to run iterative applications if the amount of memory used for cached operations is close to the actual available memory in the cluster environment. In that case, the reason for the poor results in performance is because larger amount of memory participates in the caching operation and in turn, only a small amount memory is available for computing operations of actual algorithms. Stratosphere shows favorable characteristics as a general parallel computing framework, but it has no support for iterative algorithms and spends more computing resources than Spark for the same amount of work. In another aspect, applications based on Stratosphere can achieve benefits by manually setting compiler hints when developing the code, whereas Spark has no corresponding functionality.

Regression discontinuity designs are increasingly popular quasi-experimental research designs among applied econometricians desiring to make causal inferences on the local effect of a treatment, intervention, or policy. They are also widely used in social, behavioral, and natural sciences. Much of the existing literature relies on the assumption that the discontinuity point or cutoff is known a-priori, which may not always hold. This thesis seeks to extend the applicability of regression discontinuity designs by proposing a new approach towards detection of an unknown discontinuity point using structural-break detection and machine learning methods. The approach is evaluated on both simulated and real data. Estimation and inference based on estimating the cutoff following this approach are compared to the counterfactual scenario where the cutoff is known. Monte Carlo simulations show that the empirical false-detection and true-detection probabilities of the proposed procedure are generally satisfactory. Finally, the approach is further illustrated with an empirical application. / Graduate

Les noyaux riches en neutrons de masse A=100-110 constituent une région de grand intérêt pour l'étude de la structure nucléaire loin de la stabilité. De précédentes études de cette région de masse ont déjà révélé la complexité de l'évolution de la collectivité et de la déformation dans les chaînes isotopiques de Zr, Mo, Ru et Pd. Afin d'étendre les données expérimentales sur la collectivité à des états de plus haut spin et à des noyaux plus riches en neutrons, des temps de vie d'états excités ont été mesurés dans des noyaux produits par une réaction de fusion-fission en cinématique inverse au GANIL. Les fragments de fission étaient séparés et identifiés en A et Z grâce au spectromètre magnétique de grande acceptance VAMOS tandis que le rayonnement gamma était détecté dans l'ensemble de détecteurs germanium EXOGAM. Environ vingt temps de vie d'états 2+, 4+ et 6+ ont été extraits à l'aide du plunger de Cologne. Cette expérience représente la première mesure RDDS dans des fragments de fission identifiés évènement par évènement à la fois en A et Z.Cette étude des noyaux de masse A=100-110 est complétée par des calculs auto-cohérents de champ moyen et au-delà avec la force de Gogny (D1S). La structure des états fondamentaux et excités est décrite dans le cadre du modèle de Hartree-Fock-Bogoliubov avec des contraintes sur les déformations axiale et triaxiale. Les excitations individuelles sont étudiées par des calculs bloqués et les états de haut spin sont décrits dans l'approximation du champ tournant. Enfin, la méthode de la coordonnée génératrice approchée par un hamiltonien collectif en cinq dimensions (5DCH) est appliquée aux états collectifs de basse énergie. Les résultats sont comparés aux mesures de la collectivité. / Neutron-rich nuclei of mass A=100-110 are of great interest for the study of nuclear structure far from stability. Previous experimental and theoretical studies suggest a complex evolution of deformation and collectivity in the isotopic chains of Zr, Mo, Ru and Pd.In order to extend information on the evolution of the collectivity towards higher spin states and more neutron-rich nuclei, lifetimes of excited states were measured in nuclei produced through a fusion-fission reaction in inverse kinematic at GANIL. Fission fragments were separated and identified in both A and Z with the high acceptance magnetic spectrometer VAMOS while the EXOGAM germanium detectors array was used for the coincident gamma-ray detection. Lifetimes of about twenty excited states were extracted using the plunger device of Cologne. This is the first RDDS measurement on fission fragments which are identified in A and Z on an event-by-event basis. The study of this mass region is completed by theoretical calculations using self consistent mean field and beyond mean field methods implemented with the Gogny force (D1S). The structure of the ground states and the excited states is described with Hartree-Fock-Bogoliubov calculations with constraints placed on the axial and triaxial deformations. Individual excitations are investigated through blocking calculations and the high spin states are studied through cranking calculations. Finally, an approximated generator coordinate method (GCM+GOA) using the 5DCH hamiltonian is used to describe the low energy collective states and to interpret the experimental evolution of the collectivity.