Esta tese de doutoramento trata da disputa entre o PSDB e o PT nas eleições presidenciais ocorridas no Brasil entre 1994 e 2006. O objetivo principal é identificar quais são as variáveis que estão associadas ao voto nos dois partidos que nos permitem prever com alguma precisão as chances de voto em cada um deles. A análise será procedida com auxílio de bancos de dados representativos dos eleitores brasileiros. São observadas diversas variáveis, classificadas em dois grupos: no primeiro estão as variáveis relacionadas a aspectos de longo prazo que incidem sobre o comportamento eleitoral, as características demográficas, sócio-econômicas e político-ideológicas dos eleitores, no segundo estão as variáveis ligadas diretamente ao processo eleitoral, as avaliações dos governos e dos candidatos. A hipótese mais importante a ser testada é que as variáveis de curto prazo têm impacto muito maior sobre as chances de voto nos partidos do que as variáveis de longo prazo. Isso indica que o contexto eleitoral afeta mais o resultado das eleições presidenciais do que qualquer aspecto estrutural, seja social, econômico ou político. Os resultados obtidos com a utilização de regressões logísticas corroboram a hipótese e apontam para diferenças importantes entre o voto no PSDB e no PT. / This Ph.D. thesis is about the dispute between the PSDB and PT for presidential elections ran in Brazil between 1994 e 2006. Its main objective is to identify the variables linked to votes given for both parties which would lead us to predict under a certain precision degree the chances each one has against other. The analysis is based on relevant databases from brazilian electoral activities. Several variables were observed and separated in two different groups: the first contains the ones related to long term aspects linked to the voters behavior as well as demographics, socio-economic, and political-ideological variables, while the second one contains the variables directly linked to the electoral process, to candidates evaluations and the country issues. The most important hypothesis to be tested is the one of short term variables making a bigger impact over each party vote chance when compared to the long term ones. Considering the hypothesis valuable, this presents the scenary of the electoral context with a bigger effect over presidential election than any other voter structural aspect either social, economic or political. The given result corroborates with the hypothesis and came from logistics regressions, pointing out relevant differences between votes given in favour of PSDB and PT.

A malha rodoviária brasileira é caracterizada por vias com níveis de qualidade heterogêneos. Em geral, rodovias que possuem cobrança de pedágio apresentam melhor estado de conservação. Essa dissertação tem como objetivo avaliar a escolha de rota de diferentes classes de usuários, considerando-se a cobrança de pedágio e a diferença na qualidade das vias. A avaliação envolveu a realização de duas fases de pesquisa. A primeira fase buscou estimar os valores monetários dos atributos de tempo e distância de viagem através da técnica de Preferência Declarada. A segunda fase utilizou estes valores monetários como parâmetros da função de custo generalizado da modelagem de alocação de tráfego para estimar a escolha de rota dos usuários. A partir da metodologia proposta, um estudo de caso foi realizado em um trecho rodoviário que apresenta uma rodovia pedagiada e uma rota concorrente de menor qualidade sem cobrança. Os resultados da aplicação prática não refletiram os dados de fluxo observados nas rodovias avaliadas. A partir desta constatação, um questionário complementar foi elaborado buscando identificar o comportamento habitual dos usuários na região de estudo. Através dos resultados da pesquisa complementar, verificou-se que a maioria dos entrevistados não possuía conhecimento sobre a rota alternativa. A utilização dos dados provenientes destes usuários na modelagem não se mostrou apropriada, uma vez que sua escolha de rota não é baseada nas relações de tempo, distância e penalidades nas vias, consideradas nos modelos de alocação de tráfego. Com isso, pode-se verificar neste estudo que os princípios e limitações das técnicas e ferramentas utilizadas, bem como de sua utilização conjunta, devem ser avaliados. / The Brazilian road system is characterized by roads with heterogeneous level of quality. Generally, tolled roads present better quality conditions than those not tolled. This work aimed to evaluate the route choice of different user’s classes considering toll collection and roads of different qualities. This evaluation was carried out with a methodology composed of two phases preceded by definition of the study region and the users classes considered. The first phase intended to estimate monetary values of time and distance through the Stated Preference technique. The second phase used these monetary values as parameters in an assignment model generalized cost function to estimate user’s route choice. From the proposal methodology, a case study was carried out in a road stretch that presents a tolled highway and a worse quality and without charge alternative route. The practical application results did not reflected the observed flow data in the evaluated highways. From these results, a complementary questionnaire was elaborated to identify the habitual user’s behavior in the study region. It was observed that the majority of the interviewed ones did not have knowledge of the alternative route. Using these data in the modeling was not appropriate because their route choice is not based in time, distance and road penalties relationship like in the assignment models. Thus, it was verified in this study that the principles and limitation of the used techniques and tools, as well as the joint use of these, might be taken in consideration.

It is next to impossible to let a day go by without coming into contact with a computer system in some way, either by direct usage, or making a telephone call that is directed by computer systems. Due to the importance of computer systems we are all concerned with their quality, directly or indirectly, a malfunctioning system will disturb our lives in more or less drastic ways. Each stakeholder promotes his qualities and priorities for the system, and this will eventually lead to conflict, depending on relations between qualities. There is no simple solution to the problem of relating and conflicting quality attributes. But by making an informed decision it is possible to soften the effects of the trade-offs and gain a higher level of total software quality. The way to make an informed decision is to have knowledge about the relations between quality attributes along with insights of the consequences of the decision. This report gives a short introduction to the problem, and surveys both the relations stated within academia, and industry. The intention is to explore the knowledge present within academia and collect and present the established relations found by researchers. Further this report presents a survey carried out at companies working in different areas with software engineering, to show which relations that are visible to industry and how they are handling the relations between quality attributes. This gathered knowledge intends to support the ambition towards higher quality in software engineering.

There is an emerging field in the design and development of complex systems, where systems are built upon components which in themselves are large scale systems – system of systems. Among other things, the system of systems viewpoint emphasises on open complex systems. In this thesis, the model of open computational systems is used to convey the constituents, dependencies, and interactions of such complex distributed systems. These open complex systems are exposed to critical events, occurring in the systems execution environment. Moreover, these events may have negative effects on the system at hand, resulting in system behaviour diverging from intended. Also, to take all possible affecting events in consideration when designing the system is impossible. By being able to instrument the system at hand in real time, i.e. online, one may be able to compensate the effects caused by critical events. However, to enable online instrumentation, one needs a supporting methodology which handles issues of an online nature and supporting technologies. In this thesis, this support is enabled by the methodology of online engineering and the technology of visualization. Furthermore, these instrumentations can be performed by cognitive agents – both human and software – which may explore and refine a specific system in conformance with their own, or cooperative, agendas and qualitative goals. To be able to perform the instrumentation, the cognitive agents need to be able to observe the phenomenon at hand to gain situation awareness, which in itself lies as a foundation for the decision process, carried out during the instrumentation phase. With this in mind, one quickly realizes the importance of enabling observation of open computational systems for both human and software cognitive agents. If human cognitive agents are involved in applying the methodology; the requirements on how the system is represented for the observing human agent – how the system is visualized – grows even more important. In this thesis, we emphasise visualization technology as a supporting technology for human cognitive agents in their observation process. By providing human cognitive agents with visualization technology, we may enhance the result of their observation process and thereby also increase the possibility to reach their qualitative goals. Hence, visualization of open computational systems affects a human cognitive agent’s situation awareness, which in itself lies as the foundation for the decision making process on instrumentation of the specific system at hand in conformance with the agent’s qualitative goals. This thesis will present an evaluation of a supporting tool for visualization of systemic qualities in open computational systems. Such tool must supply functions which convey the set of requirements put forward by the selected model, method, and technology. Moreover, the evaluation will be accompanied by appropriate recommendations for improvement of such a tool.

The notion of airplane stability and control being a balancing act between stability and control has been around as long as aeronautics. The Wright brothers’ first successful flights were born of the debate, and were successful at least in part because they spent considerable time teaching themselves how to control their otherwise unstable airplane. This thesis covers four aspects of handling for large transport aircraft: large size and the accompanying low frequency dynamics, the way in which lifting surfaces and control system elements are modelled in flight dynamics analyses, the cockpit feel characteristics and details of how pilots interact with them, and the dynamic instability associated with Pilot Induced Oscillations. The dynamics associated with large transport aircraft are reviewed from the perspective of pilot-in-the-loop handling qualities, including the effects of relaxing static stability in pursuit of performance. Areas in which current design requirements are incomplete are highlighted. Issues with modelling of dynamic elements which are between the pilot’s fingers and the airplane response are illuminated and recommendations are made. Cockpit feel characteristics are examined in detail, in particular, the nonlinear elements of friction and breakout forces. Three piloted simulation experiments are described and the results reviewed. Each was very different in nature, and all were designed to evaluate linear and nonlinear elements of the cockpit feel characteristics from the pilot’s point of view. These included understanding the pilot’s ability to precisely control the manipulator itself, the pilot’s ability to command the flight path, and neuro-muscular modelling to gain a deeper understanding of the range of characteristics pilots can adapt to and why. Based on the data collected and analyzed, conclusions are drawn and recommendations are made. Finally, a novel and unique PIO prediction criterion is developed, which is based on control-theoretic constructs. This criterion identifies unique signatures in the dynamic response of the airplane to predict the onset of instability.

Les ressources techniques, tactiques, physiques et mentales sont autant d’éléments à prendre en compte pour accéder à la performance en compétition. Pour autant, la spécificité du rugby à 7 attribue à la dimension physique une place particulière. La planification de l’entraînement, et en particulier la façon de gérer la charge à l’approche d’une compétition, devient ainsi une préoccupation permanente. Le but de cette thèse était d’étudier la réponse de joueurs internationaux de rugby à 7 à une gestion stratégique de la charge d’entraînement à l’approche d’une compétition. Notre première étude a consisté à décrire la distribution de la charge de travail sur une saison et caractériser différents modèles de périodisation. Ce travail a permis d’identifier l’affûtage comme une stratégie largement utilisée avant chaque tournée pour optimiser la performance physique des joueurs. Notre deuxième étude a consisté à montrer les effets d’une période d’affûtage sur les qualités physiques. L’observation des cinétiques de surcompensation des qualités de force, de vitesse et de répétition de sprints a démontré qu’une fenêtre comprise entre 9 et 16 jours constituait une durée optimale pour maximiser les gains. Enfin, notre dernière étude a porté sur une stratégie de pré-conditionnement mise en œuvre le jour de la compétition. Il s’agissait d’expérimenter les effets d’une séance d’entraînement, réalisée 2 heures avant le premier match du tournoi, sur les réponses physiques et psycho-physiologiques des joueurs. La capacité à répéter des sprints et l’activité physique en match n’ont pas été améliorées mais l’état psychophysiologique a été positivement influencé par cette séance. A travers cette thèse, nous avons donc essayé de dessiner une image générale d’une saison internationale de rugby à 7 et de questionner la pertinence de stratégies de périodisation mises en œuvre par les entraîneurs de l’équipe de France à l’approche des compétitions. Les conclusions scientifiques issues de ces études doivent être appréhendées comme des orientations qui guident les entraîneurs dans la construction de leurs scénarios d’entraînement. / The technical, tactical, physical and mental resources of a team are key factors to access to the performance in competition, and the Rugby-7’s specificity attributes a particular place to the physical dimension. Training periodization and in particular the way to manage training load prior to competition, is a permanent concern. The aim of this thesis was to study the response of international Rugby-7s players to strategic training load management in the pre-competition phases. Our first study examined the workload distribution over a season and identified different periodization models. This work helped identify the tapering phase as a strategy widely used before each competition to optimize players’ physical performance. The second study investigated the effects of a tapering period on physical qualities. The supercompensation kinetics of strength, speed and repetition of sprints qualities showed that a window between 9 and 16 days was an optimal duration to maximize physical performance. The final study focused on a preconditioning strategy implemented on the day of the competition. The aim was to examine the effects of a morning training session, completed 2 hours before the first match of the tournament, on the physical and psycho-physiological responses of the players. The repeated sprinting ability and physical activity in match was not improved, but the psychophysiological state was positively influenced by this session. Through this thesis, a general picture of a Rugby-7s international season was drawn and the relevance of periodization strategies implemented by the French team coaches in the approach of competitions was questioned. The scientific findings from these studies could be used as guidelines for coaches in building their training programs.

Research shows the professional benefits of strong mentoring relationships. However, the majority of the extant mentoring research is either mentor-centric or Western-oriented. Even with the increasing trend of protégé-centric research, relatively little research exists that examines mentoring in India and what protégés from that country would most desire from the mentoring relationship and their mentors. This chapter uses the countries in the emerging economies group represented by the acronym BRIC (Brazil, Russia, India, China), combined with Hofstede’s classification system for cross-cultural comparisons. Findings from mentoring studies from countries economically and culturally similar to India are then generalized to suggest criteria Indian protégés would find most appealing in their mentors. Finally, the chapter concludes with a discussion of future areas for research on mentoring in India.

Desired qualities of Mobility-as-a-Service (MaaS) represent what potential users want, hope for, and expect of MaaS without having experienced the service yet. Knowledge about the desired qualities is essential to design MaaS to meet users’ needs. However, little attention has been paid to understand potential users' desired qualities in the literature on MaaS. This paper addresses this void in literature with an explorative study with potential MaaS users in Sweden. The results of the study are desired qualities of MaaS such as reliability, entertainment and inclusivity. The study contributes with a thematization and classification of desired qualities of MaaS. The themes are desired qualities that make life easier, desired qualities that provide assurance, desired qualities that reward, desired qualities that appeal to inclinations and desired qualities that maintain preferred habits. The four levels of classification are functional, emotional, life-changing, and social significance. Researchers and service designers could work backward from these desired qualities to design user-centred services towards user satisfaction.

A modern ghter aircraft has an advanced ight control system which highly augmentsthe control inputs from the pilot. To verify a new iteration of the control system is a timeconsuming and expensive task. It is desired to nd qualities that is not satisfactory to thepilot as early as possible in the verication process to reduce the cost for design changes.The primary objective of this thesis is to develop methods that can be used for automaticalevaluation of aircraft ying qualities from the data provided by a desktop simulator. A desktopsimulator is cheap to use compared to ight tests and tests with a pilot in a simulator.Only ghter aircraft in the precision ight phase are studied however the methods developedcould easily be extended to include other types of aircraft and other phases of ight.To evaluate the ying qualities two sets of criteria are used the MIL-F-8785C standardand the Gibson criteria. The MIL-F-8785C standard uses a second order linear system toevaluate the aircraft's ying qualities. The linear system is estimated from the nonlineardata and evaluated against the MIL-F-8785C standard. The Gibson criteria studies the timeand frequency domain directly and are designed to work with highly augmented aircraft.The set of Gibson criteria used in this thesis primary evaluates data from the time domainhowever one criterion from the frequency domain is studied.The methods developed to evaluate the ying qualities from the MIL-F-8785C standardonly works for a small part of the ight envelope furthermore they show a large dierencefor what is considered acceptable ying qualities. Because of this the methods developed forthe MIL-F-8785C standard are considered not to be suited for evaluating ying qualities forhighly augmented aircraft. The methods developed to evaluate the ying qualities againstthe Gibson criteria works for a large part of the ight and also show a high accuracy. Thismakes the methods suited for evaluation of the ying qualities.

A review of the available research indicates that relatively little is known about how districts employ Academic Achievement Teams or similar mechanisms to reduce declines in student achievement and sustain increased student achievement at the elementary school level (Kutash, Nico, Gorin, Rahmatullah, & Tallant, 2010).   Turning around chronically low-performing schools is challenging work requiring a systemic rather than school-by-school approach (Robinson & Buntrock, 2011). The most successful turnaround efforts have both high-impact leaders and the district capacity to initiate, support, and enhance transformational change through the use of data. Educational leaders on all levels are realizing meaningful information can only be acquired through a proper analysis of data and good decisions are based on this thoughtful process of inquiry and analysis (Creighton, 2007).  The intent of this study was to identify practices of Academic Achievement Teams that facilitated student achievement.  Interviews were conducted with principals, directors of elementary education, a teacher, and district liaison representing the Virginia Department of Education\'s Office of School Improvement to gain insight into the operational and organizational structures of the Academic Achievement Teams. A qualitative design was selected to conduct this descriptive cross case study. In addition to the one-on-one interviews, observations of the Academic Achievement Team meetings and review of documents from each of the two study schools were examined to gain additional perspective regarding how the Academic Achievement Team operated to increase student achievement.  The interviews, observations, and document reviews were analyzed using the Constant Comparative Method to understand the specific practices employed by Academic Achievement Teams that increased student achievement at two elementary schools (Maykut & Morehouse, 1994). / Ed. D.