Mémoire numérisé par la Direction des bibliothèques de l'Université de Montréal.

Submitted by Rafael Santana (rafael.silvasantana@ufpe.br) on 2017-04-26T16:31:53Z No. of bitstreams: 2 license_rdf: 1232 bytes, checksum: 66e71c371cc565284e70f40736c94386 (MD5) DISSERTAÇÃO- Igor Nogueira de Oliveira.pdf: 1932477 bytes, checksum: de99161e3dad0c3f6869af876a3e2b5a (MD5) / Made available in DSpace on 2017-04-26T16:31:53Z (GMT). No. of bitstreams: 2 license_rdf: 1232 bytes, checksum: 66e71c371cc565284e70f40736c94386 (MD5) DISSERTAÇÃO- Igor Nogueira de Oliveira.pdf: 1932477 bytes, checksum: de99161e3dad0c3f6869af876a3e2b5a (MD5) Previous issue date: 2016-03-04 / Desde sua padronização, o protocolo Hypertext Transfer Protocol (HTTP) tornou-se o estado da arte em protocolos de transporte da internet, sendo utilizado para transmissão de arquivos hipertexto e hipermídias, como áudio e vídeo de forma cada vez mais interativa. Porém, o HTTP em suas versões 1.0 e 1.1 apresenta pontos que podem ser otimizados, como o fato de atender requisições de forma síncrona, o que normalmente atrasa a renderização de páginas Web podendo vir a afetar a qualidade de experiência dos usuários. Recentemente, o protocolo HTTP foi atualizado para a versão HTTP/2, recebendo diversas modificações, direcionadas principalmente a melhorias no tocante ao uso dos recursos de rede. Dentre estas melhorias, podese citar a adição do recurso push, que permite que o servidor Web responda a uma solicitação com mais de um recurso ao mesmo tempo. Este trabalho apresenta uma análise de desempenho do recurso push no transporte de páginas Web em conexões HTTP/2. Para tanto, foram realizados experimentos em um ambiente simulado, replicando características de rede presentes na internet. Foram adotados o Total Download Time (TDT) e o Page Load Time (PLT) como métricas de análise e a execução do experimento foi realizada através de requisições de páginas web com diferentes quantidades e tamanhos de objetos. Através dos resultados obtidos, foi possível observar que apesar do protocolo HTTP/2 possuir recursos para a melhoria do carregamento de páginas Web o uso inadequado destes recursos pode causar, em determinadas configurações, degradações no carregamento das páginas. / Since its standardization, the Hypertext Transfer Protocol (HTTP) has been considered the state of the art for transmission of hypertext and hypermedia files, such as audio and vídeo in an interactive way. However, 1.0 and 1.1 versions of HTTP present some weaknesses that may be optimized, such as synchronized requests, which typically slow Web pages rendering thus affecting the end user experience. Recently, the protocol was updated to the 2.0 version and received several modifications, focused mainly on improvements in the network usage. Among these improvements one can cite the addition of push, a feature that allows the server to reply to a request with more than one resource simultaneously. This work presents a performance analysis of push feature on the transport of web pages on HTTP/2 connections. Therefore, experiments were conducted on a prototype environment using Total Download Time (TDT) and Page Load Time (PLT) as metrics, and Web page requests with different amounts and sizes of objects. From the results, one can observe that despite presenting features to improve Web page load time, improper use of these HTTP/2 resources may lead, on certain conditions, to the opposite effect.

Complex concurrent APIs are difficult to reason about annually due to the exponential growth in the number of feasible schedules. Testing against reference solutions of these APIs is equally difficult as reference solutions implement an unknown set of allowed behaviors, and programmers have no way to directly control schedules or API internals to expose or reproduce errors. The work in this paper mechanically generates a drop-in replacement for a concurrent API from a formal specification. The specification is a guarded command system with first-order logic that is compiled into a core calculus. The term rewriting system is connected to actual C programs written against the API through lightweight wrappers in a role-based relationship with the rewriting system. The drop-in replacement supports putative what-if queries over API scenarios for behavior exploration, reproducibility for test and debug, full exhaustive search and other advanced model checking analysis methods for C programs using the API. We provide a Racket instantiation of the rewriting system with a C/Racket implementation of the role-based architecture and validate the process with an API from the Multicore Association.

Java is currently a widely used programming language. However, there is no formal definition of Java's semantics. Consequently, Java code does not have a universal meaning. This work discusses recent attempts to formalize Java and presents a new formalism of Java called Javalite. In contrast to common approaches to formalization, Javalite is purely syntactic in its definition. Syntactic operational semantics use the structure of the language to define its behavior. Javalite models most Java features with notable exceptions being threads, reflection, and interfaces. This work presents an executable semi-formal model of Javalite in PLT Redex. Being executable means that Javalite programs can be run using this model. We then render the semi-formal model in the Coq theorem prover and present theorems stating that the operational semantics are decidable and deterministic. This formal model can then be used to facilitate research in areas such as proving properties of algorithms that perform various analyses on Java code, e.g. verification, optimization, and refactoring.

This document is a thesis submitted in partial fulfilment of the requirements of the University of Hertfordshire for the degree of Doctor of Philosophy (Part Time) in 'EMC in Wire-line Communications' in the School of Electronic, Communication and Electrical Engineering at the University of Hertfordshire. It describes a programme of research into the modelling and measurement of radio frequency interference emissions from various communication networks including Power Line (Tele)communications (PLC/PLT) and Digital Subscriber Line (DSL). An introduction and literature review are followed by the results of practical measurements on installed networks. These measurements include antenna gain and Longitudinal Conversion Loss (LCL). Power line communication networks, splitterless DSL and home phoneline networks in buildings are studied and modelled and the models are compared with the measured results. Improved EMC test methods are also described, in particular the modelling and design of four types of portable antennas for use in radiated EMC measurements with improved sensitivity at frequencies up to 30 MHz. The first type is a set of three manually tuned loop antennas covering 100 kHz - 30 MHz. The second is a set of three loop antennas that cover a similar frequency range but with remote tuning via an optical fibre link, under the control of software which also controls an EMC measuring receiver. The third type is a larger (1.6 m diameter) tuned loop covering 1.75 - 10 MHz that allows the measuring system noise floor to be below the typical atmospheric noise floor. The fourth type is an electrically short dipole covering 10 - 30 MHz with improved matching. The protection requirements for various types of radio communication services are analysed and are compared with emission levels from various types of wireline communication network. A review of existing applicable EMC standards and standards under development is also presented.

Measurement collection is a primary step towards analyzing and optimizing performance of a telecommunication service. With an Mobile Broadband (MBB) network, the measurement process has not only to track the network's Quality of Service (QoS) features but also to asses a user's perspective about its service performance. The later requirement leads to "user-side measurements" which assist in discovery of performance issues that makes a user of a service unsatisfied and finally switch to another network. User-side measurements also serve as first-hand survey of the problem domain. In this thesis, we exhibit the potential in the measurements collected at network edge by considering two well-known approaches namely crowdsourced and distributed testbed-based measurements. Primary focus is on exploiting crowdsourced measurements while dealing with the challenges associated with it. These challenges consist of differences in sampling densities at different parts of the region, skewed and non-uniform measurement layouts, inaccuracy in sampling locations, differences in RSS readings due to device-diversity and other non-ideal measurement sampling characteristics. In presence of heterogeneous characteristics of the user-side measurements we propose how to accurately detect mobile coverage holes, to devise sample selection process so to generate a reliable radio map with reduced sample cost, and to identify cellular infrastructure at places where the information is not public. Finally, the thesis unveils potential of a distributed measurement test-bed in retrieving performance features from domains including user's context, service content and network features, and understanding impact from these features upon the MBB service at the application layer. By taking web-browsing as a case study, it further presents an objective web-browsing Quality of Experience (QoE) model.

This project was undertaken to analyze workshop evaluation forms completed by just over 2000 participants who attended workshops in 1995 throughout California related to the Project Learning Tree (PLT) environmental education program. These workshops were designed for teachers and other educators working with students from preschool through eighth grade.

Memoria para optar al título de Ingeniera Civil de Minas / El presente trabajo se enmarca en un convenio entre el laboratorio ALGES de la Universidad de Chile, en conjunto con la empresa BHP Billiton. Los datos utilizados corresponden a Minera Escondida Ltda., operada por BHP Billiton. El objetivo principal es la elaboración de un modelo, mediante cokriging, para el sector Escondida , en el que se caracterice el parámetro dureza , considerando las tres variables asociadas: SAG Power Index (SPI), Work Index de Bond (BWi) y Ensayo de Carga Puntual (PLT, medido por el índice Is50). Por una parte, se tiene el modelo de bloques, con información de atributos geológicos. Por otro lado, se tienen los datos de dureza. Existen dos sets de datos: uno corresponde a los índices SPI y BWi, tomados sobre un soporte de 16m; y el otro conjunto de datos corresponde a resultados de PLT, en soporte de 1,5m, y con una densidad de muestreo mayor en profundidad. Las unidades de estimación para Escondida se definen según las alteraciones, que son tres: Cuarzo-sericita-arcilla (1-QSC), Clorita-sericita-arcilla (2-SCC) y Potásica (3). Al estudiar la correlación de la variable PLT con las otras dos, los resultados son desfavorables, lo que se ve acentuado por la existencia de datos muy extremos y discontinuidades a pequeña escala, además de ser esperable dado que el ensayo PLT mide respuesta a la compresión, mientras que SPI y BWi miden impacto y abrasión. Se realiza entonces, un cokriging sólo con SPI y BWi, que presentan buena correlación entre sí. PLT se aplica sólo para definir dos dominios, según un valor umbral de 1,8; debido a que en la planta de proceso se ha determinado que esta distinción es importante. El cokriging se efectúa en base a los datos residuales, los que se calculan sustrayendo la media local, obtenida por kriging de tendencia, a los datos. Se aplica cokriging simple con media cero. Una vez estimados los residuos, se suma la media en cada nodo y se obtiene la estimación final. La validación del modelo obtenido se hace mediante comparación estadística, gráficos de deriva espacial, y visualizaciones 2D. Se ve que la diferencia entre la media de los bloques estimados y la media de los datos, en general es más baja para el dominio PLT<1,8 . La menor precisión de los resultados con respecto a los datos para PLT mayor , se ve también en los gráficos de deriva. Esto puede deberse a que los datos de PLT menores y mayores al umbral están bastante entremezclados, por lo que finalmente hay varios datos de PLT mayor que quedan dentro del dominio PLT menor . Por el contrario, la tendencia espacial de las variables estimadas es fiel a la de los datos para el dominio de menor PLT. No obstante, en todos los casos se ve que la tendencia a lo largo del eje Z logra ajustarse bastante bien. La principal conclusión, es que los modelos estimados permiten tener una clara noción de cómo varían en el espacio los índices de dureza SPI y BWi, en especial a través de Z. Sin embargo, los resultados son menos confiables en profundidad, dado que, en esta zona, reproducen fundamentalmente los modelos de tendencia. Por otra parte, se pierde información relativa a valores extremos, en especial para SPI, pues el cokriging a partir de residuos tiende a concentrar los resultados en torno a la media. La principal ventaja del trabajo realizado reside en la abundante cantidad de datos; sin embargo, esto se contrasta con poca información en lo que respecta a la caracterización geológica y la comprensión del proceso en la planta. / Laboratorio ALGES Universidad de Chile - BHP Billiton

Le trafic membranaire est un processus universel qui est essentiel pour la fonction neuronale dans un large spectre de fonctions. De la croissance neuronale et le développement morphologique à la libération des neurotransmetteurs et la plasticité synaptique, il prend en charge l'activité neuronale et donne d'innombrables questions qui animent la recherche sur la neurobiologie d'aujourd'hui. Notamment, l’exocytose des endosomes de recyclage (ER) dans les compartiments somatodendritiques participe à la transmission synaptique et à la potentialisation synaptique à long terme (PLT). Cependant la machine moléculaire sous-tendant l’exocytose des ER reste encore méconnue. Afin d’identifier les protéines SNAREs vésiculaires (v-SNARE) impliquées dans les différentes formes d’exocytose des ER postsynaptiques, nous avons d'abord imagé les protéines VAMP neuronales fusionnées avec la pHluorine, une GFP mutée sensible au pH dans les neurones de l’hippocampe en culture. Nous avons constaté que seulement VAMP2 et VAMP4, mais pas VAMP7, rapportaient des événements d’exocytose somatodendritique dans les neurones matures. Après avoir identifié ces deux protéines candidates, nous avons utilisé la combinaison de différentes techniques de régulation négative chronique ou aiguë pour désactiver leur fonction et observer les conséquences sur l’exocytose des ER, la transmission synaptique basale ou la PLT. Nos résultats suggèrent que VAMP2 est impliqué dans une forme d’exocytose régulée importante pour la PLT, mais pas l’exocytose constitutive des récepteurs AMPA, qui stabilise la transmission basale. VAMP4 est nécessaire pour l'exocytose constitutive d'une grande partie des endosomes, mais l'implication fonctionnelle de ces endosomes doit encore être explorée, car la régulation négative de VAMP4 ne modifie pas la transmission basale. / Membrane trafficking is a universal process that is essential for neuronal function in a wide spectrum of applications. From neuronal growth and morphological development to neurotransmitter release and synaptic plasticity, it supports neuronal activity and gives countless questions that drive today’s neurobiology research. Notably, the trafficking of recycling endosomes (REs) in somatodendritic compartments participates in synaptic transmission and plasticity, such as long-term synaptic potentiation (LTP). However, the fusion machinery mediating RE exocytosis is still unclear. To identify the vesicular SNAREs (v-SNAREs) involved in different forms of postsynaptic RE exocytosis, we first imaged neuronal VAMP proteins fused with pH-sensitive pHluorin in cultured hippocampal neurons, and found that only VAMP2 and VAMP4, but not VAMP7, underwent somatodendritic exocytosis in mature neurons. After identifying these two candidate proteins, we used a combination of different downregulation techniques to chronically or acutely deactivate their function and observe consequences on REs exocytosis, basal synaptic transmission and LTP. Our results suggest that VAMP2 is involved in activity-regulated exocytosis important for LTP, but not constitutive postsynaptic AMPARs exocytosis, supporting basal transmission. VAMP4 is required for constitutive exocytosis of at least a large proportion of REs, but the functional implication of these endosomes still need to be explored, as VAMP4 downregulation did not alter basal synaptic transmission.

Le système endocannabinoïde est un système neuromodulateur majeur du cerveau. Ainsi, il contrôle la mémoire et l’apprentissage, et ce, principalement par l'intermédiaire des récepteurs aux cannabinoïdes de type 1 (CB1) qui régulent de manière fine les activités neuronales et gliales. Dans l’hippocampe, une communication bidirectionnelle entre neurones et astrocytes modèle la plasticité synaptique et le comportement. Il a été rapporté que les effets disruptifs des cannabinoïdes sur la plasticité synaptique et la mémoire de travail sont dépendants de récepteurs CB1 présents dans les astrocytes. Cependant, le rôle de ce récepteur dans la modulation physiologique des processus mnésiques n’est pas encore connu. De précédentes études ont également montré que les récepteurs CB1 exprimés dans les cellules hébergeant le récepteur dopaminergique D1 sont impliqués dans la modulation hippocampique de la mémoire associée aux évènements aversifs. Toutefois, leur implication dans la modulation de la formation de la mémoire associée à des évènements non aversifs ainsi que dans la plasticité synaptique sous-jacente reste encore inconnue. Dans cette thèse, mon objectif était d’identifier les mécanismes cellulaires et moléculaires par lesquels des populations distinctes de récepteurs CB1 dans des populations gliales et des régions cérébrales bien définies contribuent à la modulation physiologique de la plasticité synaptique, de l’apprentissage et de la mémoire. Pour ce faire, nous avons utilisé des souris mutantes conditionnelles dans lesquelles le récepteur CB1 a été rendu silencieux sélectivement dans les astrocytes ou dans les cellules exprimant le récepteur D1. En couplant ces modèles génétiques murins avec des approches comportementales, pharmacologiques et électrophysiologiques in vitro et in vivo, nous avons disséqué le rôle de ces populations de récepteurs CB1 dans la formation de la mémoire. Tout d’abord, nous avons montré que les récepteurs CB1 astrogliaux dans l’hippocampe contrôlaient la potentialisation à long terme (PLT) de la transmission synaptique CA3-CA1 et la mémoire de reconnaissance à long terme. En contrôlant, via la gliotransmission, la disponibilité effective de D-sérine aux récepteurs NMDA, les astrocytes sont des éléments importants contrôlant les interactions glie-neurones qui sous-tendent la plasticité synaptique et les fonctions mnésiques. Les données obtenues montrent que les récepteurs CB1 astrogliaux contrôlent la plasticité et la mémoire en régulant la disponibilité synaptique de la D-sérine aux récepteurs NMDA. Deuxièmement, nous avons montré que les récepteurs CB1 dans les cellules exprimant le récepteur D1 contrôlaient la consolidation, mais pas l’acquisition, de nouveau souvenirs et l’augmentation de la PLT induite par l’apprentissage. Ces résultats indiquent que des populations spécifiques de cellules exprimant le récepteur CB1 modulent ces processus de manière différentielle.En conclusion, ces travaux démontrent que le système endocannabinoïde dans les astrocytes est un important modulateur de l’apprentissage et de la mémoire alors que les récepteurs CB1 dans les cellules exprimant le récepteur D1 semblent importants pour des composantes spécifiques de la formation de la mémoire. Prise dans son ensemble, cette thèse apporte des preuves fonctionnelles quant à la régulation complexe de la mémoire de reconnaissance à long-terme par des populations distinctes de récepteurs CB1. / The endocannabinoid system is a major brain modulatory system that controls memory and learning mainly via the cannabinoid receptor type 1 (CB1)-dependent regulation of neuronal and glial activity. In the hippocampus, bidirectional communication between neurons and astrocytes shapes synaptic plasticity and behavior. CB1 receptors have been shown to be present in the astrocytes and to mediate the disruptive effects of cannabinoids in synaptic plasticity and working memory. Yet, it is not currently known the role of this receptor in the physiological modulation of memory processes. Also, previous studies have shown that CB1 receptors expressed in dopamine D1 receptor-expressing cells are involved in the modulation of hippocampal-dependent aversive memories. However, their involvement in the modulation of non-aversive long-term memory formation and synaptic plasticity is presently unknown. In this thesis, I aimed at identifying the cellular and molecular mechanisms by which specific CB1 receptors in distinct brain neuronal and glial populations contribute to the physiological modulation of synaptic plasticity and learning and memory. For this aim we used conditional genetic mutant mice lacking CB1 receptors specifically in astrocytes or in D1-positive cells. By coupling these genetic mouse models with behavioral, pharmacological, and in vitro and in vivo electrophysiological approaches, we dissected the role of these CB1 receptors in the formation of memory. First, we show that astroglial CB1 receptors in the hippocampus control long-term potentiation (LTP) of CA3-CA1 synaptic transmission and long-term recognition memory. By allowing physiological availability of D-serine at NMDA receptors via gliotransmission, astrocytes are important elements controlling glia-neuron interactions that underlie synaptic plasticity and memory functions. The data show that astroglial CB1 receptors control plasticity and memory by regulating the synaptic availability of D-serine at NMDA receptors. Second, we show that CB1 receptors D1-positive cells control the consolidation, but not acquisition, of new memories and the enhancement of LTP induced by learning, showing that specific subpopulations CB1 receptor-expressing cells differentially modulate these processes.Overall, by showing that the endocannabinoid system in astrocytes is an important modulator of learning and memory and by suggesting that CB1 receptors in D1-positive cells are important for specific components of memory formation, we provide functional evidence for the complex cell type-dependent regulation of long-term recognition memory by the CB1 receptors.