Em sistemas MIMO multiusuÃrio, o transmissor pode selecionar um subconjunto de antenas e/ou usuÃrios que tÃm bons canais para maximizar o rendimento do sistema usando vÃrios critÃrios de seleÃÃo. AlÃm disso, os prÃ-codificadores podem proporcionar dimensÃes livres de interferÃncia. O alinhamento de interferÃncia (IA) à baseado no conceito de prÃ-codificaÃÃo e oferece diferentes compromissos entre complexidade e desempenho. A idÃia bÃsica do Alinhamento InterferÃncia consiste em prÃ-codificar os sinais transmitidos de maneira que os mesmos sejam alinhados no receptor, em que eles constituem interferÃncia, enquanto que ao mesmo tempo os separa do sinal desejado. No entanto, a InformaÃÃo do Estado do Canal (CSI) tem sido uma preocupaÃÃo para os pesquisadores porque ela tem um impacto no desempenho de algoritmos de IA. Assim, nos propomos a analisar o desempenho da seleÃÃo de antena e diversidade multiusuÃrio em conjunto, a fim de permitir o IA oportunista usando vÃrios critÃrios com relaÃÃo à perturbaÃÃo da CSI. AnÃlises e simulaÃÃes verificam o comportamento do esquema proposto. / In multiuser MIMO systems, the transmitter can select a subset of antennas and/or users which have good channel conditions to maximize the system throughput using various selection criteria. Furthermore, precoding can provide free interference dimensions. The Interference Alignment (IA) is based on the concept of precoding and it offers different trade-offs between complexity and performance. The basic idea of Interference Alignment consists in precoding the transmitted signals such that they are aligned at the receiver where they constitute interference, while at the same time disjointed from the desired signal. However, the Channel State Information (CSI) has been a concern because it impacts the performance of IA algorithms. Hence, we propose to analyze the performance of antenna selection and multiuser diversity together in order to allow opportunistic IA using several criteria over the disturbance of CSI. Analyses and simulations verify the behavior of the proposed scheme.

O objetivo geral desta dissertação é analisar o não dito sobre os requisitos de seleção de pessoal das empresas de telefonia móvel, através de vagas encaminhadas para empresas intermediadoras de emprego. O estudo está centrado em três grandes eixos: a flexibilização, os requisitos descritos e os requisitos não descritos nas seleções realizadas por intermediadoras de emprego para as empresas de telefonia móvel. O primeiro eixo diz respeito ao contexto em que ocorrem os processos seletivos e verifica os aspectos explícitos e implícitos das relações flexíveis estabelecidas entre operadoras e empresas intermediadoras. O segundo eixo está centrado na análise dos requisitos descritos como necessários para os cargos, sendo esses critérios os anunciados para os candidatos. O terceiro eixo, por fim, analisa os requisitos não descritos, aqui nomeados como “não ditos”e não anunciados, mas praticados pelas empresas e determinantes para contratação ou exclusão de candidatos. Trata-se de uma pesquisa de abordagem qualitativo-exploratória. A coleta de dados foi realizada através de entrevistas semiestruturadas, realizadas com 14 sujeitos – gestores, selecionadores e trabalhadores – que atuam em empresas intermediadoras de emprego. O estudo constata que as vagas encaminhadas para as empresas intermediadoras são predominantemente operacionais e temporárias e que os processos seletivos realizados ocorrem mediante pressão por cumprimento de prazos e metas impostas pelas operadoras. A pesquisa revela que os requisitos de seleção de pessoal vão além dos descritos e anunciados socialmente, identificando 12 elementos que não constam no perfil formal das empresas, mas que são considerados na seleção. Com isso, questiona a negação da subjetividade e discurso de imparcialidade e neutralidade nos processos seletivos sustentados por correntes tradicionais da Administração e Psicologia. O estudo também discute a aplicação do modelo de competências no contexto de submissão e pressão por prazos em que estão inseridas as intermediadoras e conclui que o seu caráter subjetivo e ainda incipiente, em termos de ferramentas de avaliação, ajuda a mascarar os requisitos velados. / The objective of this paper in general is to analyze the untold about the requirements for personnel selection of mobile phone companies through vacancies taken to job brokerage companies. The study focuses on three main areas: flexibility, the requirements described and the requirements that are not described in the selections made by job brokerage companies to the mobile phone companies. The first axis is related to the context in which there are selective processes and it verifies the explicit and implicit aspects of flexible relations established between operators and brokerage companies. The second axis focuses on the analysis of the requirements needed for the positions described, these being the advertised criteria for applicants. The third axis, finally, discusses the requirements that are not described, here named as "untold" and not announced, but practiced by companies and fundamental to the recruitment or exclusion of applicants. This is an exploratory qualitative research approach. Data collection was conducted through semi structured interviews with 14 individuals - managers, pickers and workers - who work for job brokerage companies. The study points out that the vacancies taken to the job brokerage companies are mainly operational, temporary and the selective processes occur under pressure because of meeting deadlines and targets imposed by the operators. The research reveals that the requirements for personnel selection go beyond those described and announced socially, identifying 12 elements not found in the formal profile of the companies, but considered in the selection. Thus, it questions the denial of subjectivity and discourse of impartiality and neutrality in the selection processes supported by traditional patterns of management and psychology. The study also discusses the application of the competency model in the context of submission and pressure for deadlines in which job brokerage companies are embedded and it concludes that its subjective and still nascent nature in terms of assessment tools helps mask the hidden requirements.

Existem poucas informações sobre o uso do método do retrocruzamento visando o melhoramento de caracteres quantitativos. Este trabalho teve como objetivo avaliar a eficiência de uma geração de retrocruzamento para o desenvolvimento de linhagens superiores em soja, a partir de um cruzamento biparental. Para isso, foram utilizados como genitores os cultivares BRS-134 e EMGOPA-315, sendo este utilizado como genitor recorrente por ser mais produtivo, mais alto e mais tardio. Foram avaliadas experimentalmente 100 progênies F2:4 e 100 progênies RC1F3, em três ambientes (combinação de locais e anos): área experimental do Departamento de Genética da ESALQ/USP e Estação Experimental Anhumas, ambas localizadas em Piracicaba, SP, utilizando o delineamento látice simples duplicado 10x10 (quatro repetições) e parcelas lineares de 2 m com espaçamento de 0,5 m, contendo 30 plantas após o desbaste. Os caracteres avaliados foram: número de dias para o florescimento (DF), altura das plantas no florescimento (AF), número de dias para a maturação (DM), altura das plantas na maturação (AM), acamamento (AC) e produção de grãos (PG). Para todos os caracteres foram estimados os seguintes parâmetros: média geral ( 2:4 F e 1 3 RC F ), amplitude de variação das médias das progênies ( 2 4 F : D e 1 3 DRC F ), variância genética entre progênies ( 2 p ^ ), variância fenotípica entre médias de progênies (2 F ^ ), coeficiente de herdabilidade entre médias de progênies (h2%) x ^ e resposta esperada com seleção (Rs), para as progênies F2:4 e RC1F3. Para todos os caracteres, a média geral foi superior para as progênies RC1F3, conforme o esperado. Entretanto, para PG e AF, as variâncias genéticas foram maiores para as progênies RC1F3, bem como as amplitudes de variação das médias, o que não era esperado com base em um modelo de um loco. Para DF também houve um pequeno aumento na variância genética, embora não significativo, enquanto para DM, AM e AC a variância genética foi maior nas progênies F2:4. Devido a isso, o coeficiente de herdabilidade entre médias de progênies retrocruzadas também foi maior para PG, bem como a resposta esperada com seleção. Os resultados deste trabalho indicam, portanto, que uma geração de retrocruzamento para o genitor mais produtivo pode ser uma boa estratégia para melhorar a eficiência da seleção para produção de grãos em soja. / There is little information on using the backcross method for the improvement of quantitative traits in soybeans. This work was carried out to evaluate the efficiency of a generation of backcrossing for the development of superior soybean lines, in a population derived of a two-way cross. The cultivars BRS-134 and EMGOPA-315 were used as parents, where the last one was used as recurrent parent for being higher yielding, taller and later. One hundred F2:4 progenies and one-hundred RC1F3 progenies were evaluated in three environments (combination of locations and years): experimental area of the Department of Genetics at ESALQ/USP and Anhumas Experimental Station, both located in Piracicaba, state of São Paulo, Brazil, using a simple 10x10 lattice design twice (four replications). Plots consisted of 2 meter long spaced by 0.5 meter, with 30 plants after thinning. The following traits were evaluated: number of days to flowering (DF), plant height at flowering (AF), number of days to maturity (DM), plant height at maturity (AM), lodging (AC) and grain yield (PG). The following parameters were estimated for both, F2:4 and RC1F3 progenies: general mean ( 2:4 F and 1 3 RC F ), amplitude of the individual progeny means ( 2 4 F : D and 1 3 DRC F ), genetic variance among progenies ( 2 p ^ ), phenotypic variance on a progeny mean basis (2 F ^ ), heritability coefficient on a progeny mean basis (h2%) x ^ and expected response to selection (Rs). For all traits general mean was higher for RC1F3 progenies, as expected. However, for PG and AF, genetic variances were higher for RC1F3 progenies, as well as the amplitude of individual progeny means, which are not expected, based on a onelocus model. For DF, there was also a small increase in genetic variance, although not significant, while for DM, AM and AC genetic variances were higher for F2:4 progenies. Consequently, heritability based on a progeny mean basis was higher for PG in the backcross population, as well as the expected response to selection. General results indicate that one generation of backcross for the higher yielding parent could be a good strategy to improve the efficiency of selection for grain yield in soybeans.

D.Ed. (Educational Management) / Please refer to full text to view abstract

An overview of the literature indicates that there is a real and pressing need to explore the concept of adverse impact in more detail. The Employment Equity Act of 1998 prioritises the issue of group representivity in staff selection and although there are merits in enforcing an Affirmative Action policy, an objective assessment of the inherent requirements of job applicants is still a constitutional and procedural outcome. There is no simple mechanism that can be used to determine the manifestation of adverse impact. However, evidence thereof can be most useful in securing fairness in selection and employment decisions. There is a distinct need to examine the matter of adverse impact and its implications in the South African context. South Africa is confronted with a labour force that is growing dramatically and changing significantly. The demographic trends in our dynamically changing labour markets are likely to amplify skill and ability differences between the resulting contradictory goals of equal opportunity based on individual merit and equal employment results for subgroups of populations. Three selection strategies of a communications company were used as a basis for the assessment of adverse impact. The total population of the Call Centre selection strategy was 150. Seven instruments and measures were utilised in the study, namely, a job analysis, pre-screening exercise, psychometric ability tests, role-play simulation, structured interview, job compatibility questionnaire and a performance evaluation. The total population of the Operator Services selection strategy was 139. Five psychometric instruments were utilised in the study. The total population of the Female Leadership Development Programme selection strategy was 471. Three psychometric instruments were utilised in the study. The Adverse Impact Model was applied to each of the aforementioned strategies. In terms of the definition, adverse impact exists if the selection ratio associated with a particular passing score on a test for one sub-group of job applicants is less than 4/5th’s or 80% of the selection ration for the largest sub-group of applicants. 13 The results of the study indicate that adverse impact occurred in the following instances. In the Call Centre selection strategy, there was evidence of adverse impact in the racial analysis in the ability testing stage. There was also evidence of adverse impact in the gender analysis during this stage. There was no evidence of adverse impact in either the racial or gender analyses during the interview stage. There was no evidence of adverse impact in the racial analysis during the appointment stage. There was however evidence of adverse impact in the gender analysis during the appointment stage. The Female Leadership Development Programme selection strategy revealed evidence of adverse impact in the racial analysis during the appointment stage. The Operator Services selection strategy revealed an adverse impact in both the racial and gender analyses during the testing stage. The issue of adverse impact is obviously controversial. However, data obtained from this process will provide for informative analysis.

The selection of variables problem is ussual problem of statistical analysis. Solving this problem via Bayesian statistic become popular in 1990s. We re- view classical methods for bayesian variable selection methods and set a common framework for them. Indicator model selection methods and adaptive shrinkage methods for normal linear model are covered. Main benefit of this work is incorporating Bayesian theory and Markov Chain Monte Carlo theory (MCMC). All derivations needed for MCMC algorithms is provided. Afterward the methods are apllied on simulated and real data. 1

Les populations de rongeurs sont responsables de nombreux problèmes économiques et sanitaires. Pour ces raisons certains rongeurs sont qualifiés de nuisibles, tels que le rat brun, le rat noir et la souris domestique. En conséquence, leurs populations doivent être gérées. Cette gestion implique des mesures sanitaires et écologiques strictement indispensables mais passe aussi par l'utilisation de produits chimiques toxiques. Dans le cadre de la lutte chimique, les AVKs sont actuellement les molécules les plus utilisées. Ces AVKs inhibent la coagulation sanguine et sont à l'origine d'hémorragies mortelles. La cible pharmacologique des AVK est l'enzyme VKOR, codée par le gène vkorc1. Cependant, l'utilisation intensive des AVKs depuis les années 1950 a eu pour conséquence, chez les rongeurs, la sélection de résistance à ces molécules. La résistance de cible a été bien étudiée chez le rat brun mais est très peu connue chez la souris domestique et le rat noir / Rodent populations are responsible for many economic and health problems. For these reasons different rodents species are called pests and particularly brown rats, house mice and black rats. Their populations need to be managed. This management involves health and environmental measures strictly necessary but is also based on the use of toxic chemicals. As part of chemical control, the AVKs are currently the most commonly used molecules. These AVKs inhibit blood coagulation and are the cause of fatal bleeding. The pharmacological target of AVK is the VKOR enzymatic activity, encoded by the vkorc1 gene. However, the intensive use of AVKs since the 1950s has led to the development of resistance to these molecules. The target resistance has been well studied in the brown rat, but little is known in the house mouse and the black rat

Directional selection is prevalent in nature yet phenotypes tend to remain relatively constant, suggesting a limit to trait evolution. The genetic basis of evolutionary limits in unmanipulated populations, however, is generally not known. Given widespread pleiotropy, opposing selection on a focal trait may arise from the effects of the underlying alleles on other fitness components, generating net stabilizing selection on trait genetic variance and thus limiting evolution. Here, I look for the signature of stabilizing selection for a suite of cuticular hydrocarbons (CHCs) in Drosophila serrata. Despite strong directional sexual selection on CHCs, genetic variance differed between high and low fitness individuals and was greater among the low fitness males for seven of eight CHCs. Univariate tests of a difference in genetic variance were non-significant but have low power. My results implicate stabilizing selection, arising through pleiotropy, in generating a genetic limit to the evolution of CHCs in this species.

Studies assessing the evolution of plant traits frequently focus on pollinators as the primary drivers of floral trait evolution. However, herbivores can also play a role, and, under some circumstances, may even impose stronger selection on floral display than pollinators. This is especially true when the traits under selection are linked to anti-herbivore defense. Here I describe a study in which I quantified the relative role of herbivores and pollinators in selection for floral traits in the North American invasive plant, purple loosestrife (Lythrum salicaria). Because L. salicaria responds to leaf-chewing herbivores by producing compensatory tissue growth, and this in turn alters the architecture of the floral display, I further tested whether herbivores can indirectly modify pollinator-mediated selection through this pathway. Using a split-plot design, I measured pollen limitation and reproductive output in experimentally manipulated plants in the presence and absence of simulated herbivore damage in order to quantify the effects of damage and pollination on natural selection for floral display. My results showed that damage significantly increased direct selection (beta-i) for earlier flowering time and decreased selection on the number of inflorescences, even more than pollinators did. Because damage did not modify pollinator-mediated selection for floral display traits, the selection imposed by herbivores is likely only having direct effects. My findings demonstrate the importance of considering multiple agents of selection and their potential interactions when quantifying natural selection in a study system. In particular, it is important to consider that the agent most frequently studied may not be imposing the brunt of selection.

Un service Web est un moyen d'offrir des fonctionnalités sur un réseau en utilisant des normes ouvertes pour la description et l'invocation. Les services Web représentent une réalisation importante de l'Architecture Orientée Service (AOS), à l'aide de qui, les applications peuvent être développées rapidement avec un coût bas par couplage faible les services sur un réseau. Cela nécessite la découverte et la composition des ensembles de services interopérables, selon certaines exigences fonctionnelles et non fonctionnelles. Les services Web confrontent de nombreux défis quant à leur découverte et sélection, en raison de plusieurs facteurs tels que: le nombre important de services, le manque de registres publics capables d'offrir des mécanismes efficaces de récupération de service, de leur nature dynamique qui impose divers aspects de QoS tels que la disponibilité, le temps de réponse, etc, et le manque de sémantique dans leurs descriptions d'interface. Dans cette thèse, nous avons deux objectifs principaux. Notre premier objectif est de faciliter la sélection des services Web et d'assurer la continuité du service dans des compositions de services Web. Par conséquent, nous proposons une approche basée sur l'analyse formelle de concepts (AFC) pour classer les services Web dans un premier temps par mots clés, puis par les valeurs de similarité entre leurs opérations. Cette classification est représentée comme un treillis de concepts qui révèle les relations entre les services, ce qui facilite la sélection d'un service nécessaire ainsi que l'identification des potentiels sauvegardes (substitutions en cas de panne). Notre deuxième objectif est de guider l'utilisateur en effectuant une sélection optimisée basée sur des plusieurs critères. Nous définissons un descripteur pour l'utilisateur qui spécifie des exigences fonctionnelles et non fonctionnelles. Dans ce descripteur, les propriétés fonctionnelles sont spécifiées comme un ensemble de mots-clés. Les propriétés non fonctionnelles représentent les niveaux attendus de QoS (bon, mauvais, moyen, ..) ainsi que la composition de services exprimée en tant que liens entre les propriétés fonctionnelles spécifiées. Afin d'atteindre cet objectif, nous proposons une approche basée sur l'analyse relationnelle de concepts (ARC) qui classifie les services Web en treillis de concepts similaires à la AFC, mais enrichis avec les propriétés non-fonctionnelles. Nous proposons également un mécanisme permettant d'interroger le concept de treillis résultant basée sur RCA, selon les exigences spécifiées dans le descripteur. Nous avons validé notre proposition en utilisant des services Web réels extraits de Service-Finder et Seekda (des moteurs de recherche de services Web). Pour l'approche basée sur la AFC, nous avons récupéré un total de 145 services Web que nous avons classés en fonction de leur fonctionnalité. Nous avons montré comment sélectionner efficacement un service offrant les fonctionnalités requises et la manière d'identifier ses sauvegardes. Pour l'approche basée sur RCA, nous avons récupéré 901 services Web que nous avons classés selon leur niveau de QoS et de composabilité. Nous avons vérifié que cette approche permet une sélection efficace des services correspondant aux exigences fonctionnelles et non fonctionnelles spécifiées. / A Web service is a way of offering functionality over a network using open standards for description and invocation. Web services represent an important realization of Service-Oriented Architectures (SOA), using which, applications can be developed rapidly with a low cost by loosely coupling services over a network. This necessitates discovering and composing sets of interoperable services, according to some functional and non-functional requirements.Web services face many challenges regarding their discovery and selection, due to several factors like: the fairly large number of services, the lack of public registries capable of offering efficient service retrieval mechanisms, their dynamic nature which imposes various QoS aspects such as availability, response time, etc., and the lack of semantics in their interface descriptions.In this thesis, we have two main objectives. Our first objective is to facilitate Web service selection and assure service continuity in Web service compositions. Therefore, we propose an approach based on Formal Concept Analysis (FCA) to classify Web services first by keywords then by similarity values between their operations. This classification is represented as a concept lattice that reveals the relations between the services, which facilitates the selection of a needed service as well as the identification of its potential backups (substitution in case of failure).Our second objective is to guide the user towards performing an optimized multi-criteria based selection. We define a user requirements descriptor that specifies the needed functional and non-functional properties. Inside this descriptor, functional properties are specified as a set of keywords. Non-functional properties represent the expected QoS levels (good, bad, medium, ..) as well as the composition of services expressed as links between the specified functional properties. In order to meet this objective, we propose an approach based on Relational Concept Analysis (RCA) that classifies Web services into concept lattices similar to FCA, but enriched with the non-functional properties. We also propose a mechanism to query the resulting RCA-based concept lattices, according to the requirements specified in the descriptor.We validated our proposition using real Web services retrieved from Service-Finder and Seekda Web service search engines. For the FCA-based approach, we retrieved a total of 145 Web services that we classified by their functionality. We showed how to select efficiently a service offering the required functionality and how to identify its backups.For the RCA-based approach, we retrieved 901 Web services that we classified by their QoS and composability levels. We verified that the approach allows an efficient selection of services corresponding to the specified functional and non-functional requirements.