Le Cloud Computing est devenu l'un des grands paradigmes de l'informatique et propose de fournir les ressources informatiques sous forme de services accessibles au travers de l'Internet. Ces services sont généralement organisés selon trois types ou niveaux. On parle de modèle SPI pour “Software, Platform, Infrastructure” en anglais. De la même façon que pour les applications ``standard'', les services de Cloud doivent être capables de s'adapter de manière autonome afin de tenir compte de l'évolution de leur environnement. À ce sujet, il existe de nombreux travaux tels que ceux concernant la consolidation de serveur et l'économie d'énergie. Mais ces travaux sont généralement spécifiques à l'un des niveaux et ne tiennent pas compte des autres. Pourtant, comme l'a affirmé Kephart et al. en 2000, même s'il existe des adaptations à priori indépendantes les unes des autres, celles-ci ont un impact sur l'ensemble du système informatique dans lequel elles sont appliquées. De ce fait, une adaptation au niveau infrastructure peut avoir un impact au niveau plate-forme ou au niveau application. L'objectif de cette thèse est de fournir un support pour l'adaptation permettant de gérer celle-ci comme une problématique transverse au différents niveaux afin d'assurer la cohérence et l'efficacité de l'adaptation. Pour cela, nous proposons une abstraction capable de représenter l'ensemble des niveaux et servant de support pour la définition des reconfigurations. Cette abstraction repose sur les techniques de modèle à l'exécution (Model at Runtime en anglais) qui propose de porter les outils utilisés à la conception pour définir, valider et appliquer une nouvelle configuration pendant l'exécution du système lui-même. Afin de montrer l'utilisabilité de cette abstraction, nous présentons trois expérimentations permettant de montrer l'extensibilité et la généricité de notre solution, de montrerque l'impact sur les performances du système est faible, et de montrer que cette abstraction permet de faire de l'adaptation multiniveaux. / Cloud Computing is becoming the new paradigm for information technology to provide resources as Internet-based services. These services are basically categorized according to three layers also called SPI model (Software, Platform, Infrastructure). The same way as ``non-Cloud'' applications, Cloud services must be able to adapt themselves according to the evolution of their environment. There are many works on dynamic adaptation such as server consolidation and green computing but these works are generally specifics to one layer and do not take the others into account. However Kephart et al. have explain in 2000 that even if adaptations are, in theory, independant, they have an impact on the overall system. Consequently, an adaptation at the infrastructure layer can have an impact at the platform or at the application layers.This thesis provides an abstraction to manage adaptation as an orthogonal concern overs Cloud layers. Based on Model atRuntime (M@R) techniques which offer to use design tools to build and validate new configuration of the system at the runtime, this abstraction is able to modelize all the Cloud layers. To show the usability of this abstraction, we provide three experimentations showing the extensibility and genericity of our approach, showing that performance overhead on the system (infrastructure or platform) is weak and showing that the abstraction allows to build multi-layers adaptations.

Although Rapunzel criticism habitually concerns literary fairytales, this thesis contributes to the field a sustained examination of the feminist and patriarchal uses to which Rapunzel has been put, with close attention to the range of media, forms, and styles into which ‗Rapunzel‘ has been adapted, from 1970 onwards. It argues that each adaptation appropriates ‗Rapunzel‘ to repeat or disturb gender ideologies, and also extends or contracts the scope of the fairytale and its feminism. Underpinned by memetics, selective adaptation and fairytale theories, and Adrienne Rich‘s concept of ‗re-vision‘, individual chapters focus upon redrawing the boundaries of what makes a (feminist) Rapunzel adaptation a (feminist) Rapunzel adaptation. The thesis also examines the difficult question of why Rapunzel motifs or ‗memes‘ have persisted and whether this is due to the power of cultural ideologies or to certain universal human urges to which ‗Rapunzel‘ ostensibly appeals. As what is meant by feminism changes from the 1970s through to the present day, the selected works are considered in terms of terms of second- and third-wave feminism and postfeminism. Chapter 1 (the Introduction) establishes the approach and rationale. Chapter 2 examines the Grimm ‗Rapunzel‘ variants of 1812 and 1857 as a prelude to examining the ideological uses to which Rapunzel is put post-1970. Chapter 3 focuses on how four feminist poets subject the memes and morals of ‗Rapunzel‘ to different feminist revisions, and thereby challenge the patriarchal meanings invested by the Grimms. Chapter 4 extends this work by examining a feminist moral fable, two complex short stories, a psychological novella, and a graphic novel, in order to draw contrasts between celebratory and darker, more disturbing ‗post-fairytale‘ feminist Rapunzels. Demonstrating the many genres and media into which feminist Rapunzels have been translated, several adapters use the tale on behalf of various kinds of individualism and subjectivisation, and suggest a movement toward greater psychological complexity and interiority in their treatment of Rapunzel memes. Chapter 5 focuses on how Rapunzel memes translate to screen in the feminist reworking Rapunzel Let Down Your Hair (1978) and the postfeminist adaptations Barbie as Rapunzel (2002), Shrek the Third (2007), and Disney‘s Tangled (2010) and Into the Woods (2014). Chapter 6, the final chapter, further extends the analysis by examining Rapunzel‘s general prevalence in the cultural imagination, namely in adverts and on television. By assembling and giving fresh analyses of rare and well-known Rapunzel tales, the chapters critique the gender essentialism in fairytales and reinstate Rapunzel as key to fairytale debate. This research has led to the conclusion that post-1970s Rapunzels exemplify how fairytales appropriate or discard memes in accordance with the possibilities of genre and medium, as well as with the changing face of feminism over the last four decades.

Social-ecological systems (SES) may become “locked in” particular states or configurations due to various constraints on adaptability imposed by feedback mechanisms or by processes designed to incentivize certain behavior. While these locked-in states may be desirable and robust to disturbances over relatively short time periods, limits on system adaptations may diminish the longer-term resilience of these states, and potentially of the system itself. The agricultural SES in the Iowa-Cedar River Basin in eastern Iowa is one such system. While highly productive, culturally important, and essential to local economies, the system is facing significant economic and environmental challenges. This dissertation presents the results of a project designed to survey the adaptability of farmers in the ICRB, model their actions subject to constraints, and plot potential future states under scenarios of climate change, policy, and market conditions. We utilize a coupled agent-based model (ABM) to examine the specified resilience of the system to future climate, leveraging the ability of ABMs to integrate heterogeneous actors, dynamic couplings of natural and human systems, and processes across spatiotemporal scales. We find that farmer behavior is primarily constrained by economic factors, including federal crop insurance subsidies and the financial risk of implementing different crops or practices. Finally, we generate alternative system trajectories by modeling twenty-one scenarios, identifying actionable adaptations and pathways for transforming the system to alternative, more sustainable states.

The phenomenon of repeated evolution runs counter to expectations about the role of contingency in adaptation. However, many examples of independently acquired similar traits show that evolution sometimes does follow the same path. Factors influencing the probability of such an event include selection, trait complexity and relatedness. Previous investigations of repeated adaptation have primarily focused on low-complexity traits subject to strong selection. Studies of systems with varying levels of trait complexity, selection, and relatedness are needed to evaluate the relative contributions of these factors. The series of studies reported here 1) establishes a system for inquiry into the role of parallel adaptation among hosts and parasites and 2) provides an assessment of the role of parallel genetic change in the evolution of a complex trait. In Chapter 2, I show that all-female broods in a line of Drosophila borealis are caused by infection with a male-killing strain of Wolbachia that is very closely-related to another male-killing strain infecting a geographically and evolutionarily distant species of Drosophila. This host-parasite system, together with two other known male-killing Wolbachia strains infecting Drosophila provides a framework for investigating the role of parallel evolution in the independent acquisition of the male-killing trait among Wolbachia, as well as in the adaptation of divergent hosts to similar male-killing parasites. In Chapters 3-5, I investigate the role of parallel genetic change in a complex trait in two species of Drosophila by searching for evidence of adaptation in the Drosophila americana homologs of genes thought to underlie adaptation to climate in Drosophila melanogaster. In Chapter 3, I investigate the D. americana homolog of Alcohol dehydrogenase (Adh). In contrast with D. melanogaster, which segregates functionally distinct variants in Adh that represent local adaptation to climate, D. americana segregates little variation. This is surprising, especially because Adh of D. americana is found near a polymorphic chromosomal rearrangement that does segregate geographically-structured alleles across the species' range. In Chapter 4, I report similarities at the Phosphoglucomutase (Pgm) locus in the two species, including a shared excess of nonsynonymous variants and the presence of clinal alleles. However, while variation at Pgm of D. melanogaster is proposed to underlie local adaptation, variation at Pgm of D. americana appears to be predominantly neutral. In Chapter 5, I investigate the role of positive selection in sequence evolution in the D. americana homologs of a group of genes thought to underlie local adaptation to climate in D. melanogaster. The two species share a large geographic range and exhibit levels of sequence variation that indicate a similar effective population size, but D. melanogaster appears to undergo more frequent fixation of advantageous alleles. Approximately half of all amino acid divergence in D. melanogaster is attributable to positive selection, but I find no signs of positive selection in the investigated genes in D. americana. Overall, the results reveal little or no parallel evolution at the single genes analyzed. This lack of parallel evolution is likely a result of the high complexity of adaptation to climate as well as contingency.

Climate change and extreme weather events are affecting the environment, and people’s livelihood in both developing and developed countries. Agriculture, forestry, fishing, livestock, water resources, human health, terrestrial ecosystems, biodiversity, and coastal zones are among the major sectors impacted by these shocks. The challenge of adaptation is particularly acute in the developing countries, as poverty and resource constraints limit their capacity to act. Bangladesh fits in this category, and thus I use data from Bangladesh to analyze the adaptation process in the first and second chapter of my dissertation. In the first chapter, I investigate whether transient shocks (flood, cyclone) or permanent shocks (e.g., river erosion that leads to permanent loss of lands) have more influence on interregional migration. Findings of the study suggest that the households prefer to move to the nearest city when the environmental shock is temporary, whereas they tend to relocate over a greater distance when the environmental shock is more permanent in nature. In the second chapter, I investigate the feasibility of a set of adaptation measures to cope with hydro-climatic shocks (e.g. floods, drought, cyclones, tidal waves) and epidemic shocks (emergence or re-emergence of infectious diseases on livestock and poultry) in the agricultural sector in Bangladesh. Findings suggest that a decrease in agricultural income due to climatic and/or epidemic shocks is likely to induce households to adapt more. Developed countries are also vulnerable to extreme weather events and climatic shocks. In 2017, United States was hit by three consecutive hurricanes: Harvey, Irma, and Maria. Given the rising exposure and the increasing need to manage coastal vulnerability, the third essay focusses on understanding household preferences for financing adaptation activities in the U. S. and analyzes which mechanism, i.e., state or federal adaptation fund approach, is better suited to managing exposure to such types of natural disaster in the future.

1 / Veronica Coriano

The degree to which the normal eye can adapt to the dark is related to or dependent upon the eye's ability to regenerate visual purple. The relationship of vitamin A to the visual cycle has caused much development in improved methods of detecting vitamin A deficiency. For the most part these methods have been applied to adults and school age children. This study seeks to analyze this method as applied to pre-school age children.

In Australia the homeless population has become more diverse over the last 20 years with more young people, women and families experiencing homelessness. It is also evident that there is considerable variation in the length of time people remain homeless. How these changes relate to movements into and out of the homeless population is not well understood. This research asks: 'Is there a connection between how people become homeless, how long they remain homeless and how they 'get out' of homelessness?' A review of the literature identified two gaps directly relevant to the issue of movement in and out of homelessness. First, it is not well understood why people experience homelessness for different lengths of time when they face similar structural conditions. Second, the prevalence of substance use and mental illness reported in the homeless population has led some to conclude these factors cause homelessness. However, researchers have generally been unclear about whether such problems precede or are a consequence of homelessness. In addition, research has generally presumed a relationship between the amount of time a person is homeless and patterns of behavioural and cognitive adaptation to a homeless way of life. Yet recent research suggests that people's biographies play a significant role in the duration of homelessness. How these different findings relate to each other remains unclear. This thesis investigates these issues through a longitudinal study of homeless households. Data was gathered in two rounds of semi-structured interviews. In the first round 103 interviews were conducted. Approximately one year later 79 of these households were re-interviewed. The process of, and connections between becoming, being and exiting the homeless pathway are analysed using the 'pathways' concept. While on these pathways homeless people actively produce and reproduce social structures including both embracing and rejecting the stigma and subculture associated with homelessness. This complex world of homelessness is then analysed by extending the pathways concept by distinguishing five ideal type pathways based on the main reason for becoming homeless. They are a mental health pathway, a domestic violence pathway, a substance use pathway, a housing crisis pathway and a youth pathway. The research indicates that people on each pathway respond to the experience of homelessness differently and this has implications for the amount of time they spend in the homeless population. People on the substance use and youth pathways commonly describe themselves as 'homeless', focus on the 'here and now', use the welfare service system, are very mobile, and over time, many start to sleep rough. Their embrace of the homeless subculture commonly 'locks' them into the homeless population for long periods of time. In contrast people on the domestic violence and housing crisis pathways generally do not identify themselves as homeless and resist involvement with other homeless people. These homeless careers tend to be shorter. Then there are those who enter homelessness on the mental health pathway. They were frequently exploited in the early stages of their homeless careers and most sought to avoid exploitation by isolating themselves which then increased their marginalisation. These were the longest homeless careers. The use of the pathways concept also helps to understand how the circumstances of homeless people can change while they are homeless. The research found that some homeless people changed pathways. In particular the study found that two thirds of the people who reported substance use problems developed these problems after they became homeless. Most of these people entered the homeless population on the youth pathway. The research also found that three quarters of the people with mental health issues developed these issues after they became homeless, and that for some this was also connected to drug use. Overcoming homelessness is never easy and individuals manage the process in different ways. Again the pathways concept proved useful to understanding how homeless people accomplished this. The findings show that people travelling the different pathways require different levels and types of assistance to resolve their homelessness. The research concludes that the process of re-integration can take a long time but, given the right social and economic support, every homeless career can end.

L'échange de documents multimédia composés de plusieurs médias élémentaires tels que des vidéos, des images ou du texte, est l'une des applications les plus populaires d'Internet. Idéalement, tout usager d'Internet devrait pouvoir accéder à ces contenus et les recevoir dans un format adapté au contexte dans lequel il travaille. Un contexte utilisateur peut être défini par les caractéristiques personnelles de l'utilisateur (ex : sa langue parlée, son handicap et ses centres d'intérêt), ses préférences de présentation des contenus multimédia (ex : son lecteur multimédia préféré ou la taille d'image souhaitée), les capacités de son terminal (ex : la taille de l'écran du terminal ou les lecteurs multimédia présents) et les caractéristiques de son réseau d'accès (ex : la bande passante). Compte tenu de la combinatoire des éléments de contexte, il n'est pas envisageable de fournir autant de versions des documents multimédia que de contextes possibles : l'adaptation des contenus est donc nécessaire. L'accroissement des utilisateurs des terminaux à capacités réduites tels que les assistants personnels (par ex. PDA) exclut une adaptation côté client (ou utilisateur final). L'adaptation, côté source du document multimédia, nécessite l'implémentation de modules supplémentaires qui n'est pas toujours possible et qui peut créer une charge supplémentaire indésirable. L'adaptation par un ou plusieurs intermédiaires répond le mieux aux besoins de passage à l'échelle et d'extensibilité. Une machine intermédiaire est un nœud inséré entre le client et le serveur et dédié, par exemple, à la découverte ou à l'adaptation de services (ex : réduction de la taille d'une image ou traduction et insertion de sous-titres au sein d'une vidéo). L'intermédiation ainsi réalisée apporte une valeur ajoutée en évitant de charger l'utilisateur final et la source du document de tâches spécifiques consommatrices de ressources sans rapport direct avec le service final offert. Cette approche est celle qui est prise dans la plupart des solutions existantes. Celles-ci utilisent des intermédiaires dédiés. Il en résulte une configuration d'adaptation figée ne garantissant pas la gestion de nouvelles techniques d'adaptation (ex : les adaptations relatives à l'handicap) et ne passant pas à l'échelle. Certaines solutions, basées sur ce même modèle, intègrent l'adaptation distribuée en répartissant la charge entre les intermédiaires qui réalisent l'adaptation. Elles ne traitent cependant pas la gestion dynamique des adaptateurs qui consiste à aller chercher des adaptateurs dans le réseau, les composer et les recomposer dynamiquement en cas de disparition. Elles ne traitent pas non plus l'adaptation des documents multimédia composés qui demande un effort supplémentaire d'analyse du document et de synchronisation des médias élémentaires le composant. La première contribution de cette thèse est la conception d'une architecture appelée PAAM (pour Architecture for the Provision of AdAptable Multimedia composed documents) qui a pour but d'adapter des documents multimédia composés au contexte des usagers. L'une des originalités de cette architecture est de mettre en place une adaptation distribuée sur différents nœuds du réseau en évitant de confier l'adaptation à un serveur ou à un intermédiaire dédié. La plate-forme d'adaptation de PAAM intègre aussi bien des fournisseurs de services d'adaptation que des particuliers qui se porteraient volontaires pour exécuter des fonctions d'adaptation en donnant un peu de leurs ressources matérielles et logicielles. Les principaux éléments fonctionnels de PAAM sont : le gestionnaire du contexte utilisateur, le gestionnaire des documents multimédia composés, le planificateur et le gestionnaire d'adaptation. Le gestionnaire du contexte utilisateur et le gestionnaire des documents multimédia composés récupèrent, analysent et agrégent respectivement les informations contextuelles de l'utilisateur et les informations descriptives des documents multimédia. Le planificateur implémente un algorithme de prise de décision reposant sur des politiques d'adaptation. Ce planificateur produit un graphe d'adaptation, c'est-à-dire un ensemble d'adaptateurs organisés en parallèle ou en séquence. Ce graphe est utilisé en entrée du gestionnaire d'adaptation qui recherche ces adaptateurs là où ils se trouvent, les instancie, les compose, si nécessaire, et les recompose si un ou plusieurs adaptateurs disparaissent. Nous avons choisi d'utiliser les services Web pour implémenter PAAM afin qu'elle soit distribué, extensible, modulable, tolérante aux fautes et passant à l'échelle, répondant ainsi aux limitations des autres architectures d'adaptation. Cette solution technologique permet à PAAM de décrire des ressources d'adaptation, de les publier, de les rechercher et les instancier. Dans le cadre de la composition et de l'orchestration des services Web, nous présentons BPEL (Business Process Execution Language) et son éventuelle intégration au sein d'un gestionnaire d'adaptation pour gérer l'exécution d'un graphe d'adaptation. La seconde contribution de cette thèse est la gestion des adaptateurs (description, recherche et instanciation). Nous proposons, pour cela, une nomenclature incluant un grand nombre d'adaptateurs. Nous proposons aussi une description d'adaptateurs qui étend WSDL, et qui facilite la recherche, l'instanciation et la composition de ces ressources d'adaptation. Nous exposons par la suite le protocole de négociation et d'acceptation établi entre un gestionnaire d'adaptation et un adaptateur permettant de déterminer si cet adaptateur peut réaliser l'adaptation ou non. PAAM gérant l'adaptation distribuée sur différents nœuds du réseau, susceptibles de se déconnecter à chaque instant, nous proposons des solutions pour gérer les déconnexions dans PAAM afin de lui procurer un aspect dynamique. Afin de démontrer la faisabilité de notre architecture, nous implémentons une chaîne d'adaptation complète incluant les principales fonctionnalités de PAAM : le gestionnaire du contexte utilisateur, le gestionnaire des documents multimédia composés, le planificateur et le gestionnaire d'adaptation. Nous présentons, par la suite, une étude des coûts induits par notre implémentation de PAAM et des tests de performances qui montrent que l'utilisation des services Web n'introduit pas de surcoûts significatifs par rapport au gain obtenu en distribuant l'adaptation sur différents nœuds. Pour conclure, parce qu'elle permet de gérer une grande variété d'adaptateurs de manière distribuée, l'architecture PAAM répond bien aux limitations des architectures d'adaptation basées sur une configuration client/serveur. L'intérêt de cette approche est la possibilité d'étendre et d'enrichir le système d'adaptation et de le déployer à large échelle tout en garantissant sa robustesse.

<p>An area within psychology that looks at the strengths and positive sides of human life has emerged the last decade. It is called positive psychology and one area related to that is positive adaptation. The main purpose of this paper is to describe the natural history of females’ positive extrinsic and intrinsic adaptation from childhood to adulthood, with a focus on typical positive patterns of adaptation and how these patterns develop within the same individual. The sample consisted of about 500 Swedish girls and data were taken at age 13, 15, and 43 from the longitudinal research program Individual Development and Adaptation (IDA). Variable-oriented methods were used to study basic relationships among factors both within age and between childhood and adulthood and person-oriented methods were used to study typical patterns of adaptation and how these patterns develop, using cluster analyses and cross-tabulation of clusters. The overall results show, as expected, more distinct typical positive adaptation patterns in the intrinsic than the extrinsic area in both childhood and adulthood. Significant longitudinal developmental streams between typical positive adaptation patterns in childhood and adulthood were found and these are discussed from a dynamic system perspective suggesting the interaction between factors thru reinforcing feedback processes.</p>