RCAAP Repository

This study demonstrate the importance of Distance Education as respects on-line distance learning. It intends to discuss the possibilities of this type of learning: the credibility of distance education platforms, the use of metaphors in the creation of referentially explicit graphic interfaces, the construction of educational programs utilizing graphic libraries and efficient program languages and, further, the proposal for an on-line distance learning platform monitored by the apprentice. It concludes that the utilization of distance laboratories is a way of breaking down barriers, creating conditions that make on-line teaching an educational system that will be acceptable and will promote efficiente learning. Besides this, the proposal to create an on-line distance learning plataform monitored by the apprentice can be an important contribution toward improving the teaching-learning process.

Due to the increasing demand of consumption, the electric energy systems have faced, in the current times, problems with its operation, being the voltage control one of the most relevant matters. To provide means to solve such problems, the use of static compensators has emerged, amongst others, as a classical alternative. Although the modern solutions are mainly based on power electronic equipments, in the past, an old fashion compensator using saturated core reactor had been largely used. This is composed, basically, by a magnetic core with special windings operating in parallel with a fixed capacitor. This equipment, due to its robustness, low cost, minor maintenance requirement and good performance, consists in an attractive alternative in relation to electronic proposals when the voltage requirement control is not so strict. However, the technology of this product, appeared in the year of 60, is quite rare documented. This has created a great lack of information regarding the project and construction of these compensators. In this context, the present thesis aims at searching for mechanisms to support the design of the saturated reactor. The searched process, amongst other properties, focuses the optimization of losses associates with the operation of the device. On the basis of these procedures, in addition to computational tools utilizations to predict the equipment performance, a real size saturated reactor is designed, constructed and tested. The results are then used to validate two software (FEMM and SABER) chosen to evaluate the equipment performance analysis before the final construction.

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior

This work analyzes the technical and regulatory aspects of reactive power supply ancillary services of synchronous generators and emphasizes that Brazilian rules do not stimulate investment for such purpose and this can degrade those resources. In the past, the electrical worldwide systems used to have a centralized public utility which could include the several costs of ancillary services for electric power supply in the final price of the active power. After the privatization and decentralization, there was the need to separate the remuneration for such ancillary services. The resolutions that rule the reactive power ancillary services of synchronous generators have developed differently in several world regions. In this work it is given a description of some foreign legislations, which are then compared to the present status of Brazilian legislation, indicating its questionable aspects. Indexes are proposed for the selection of generators to supply reactive power to some electric power system area. It is proposed the public bidding for this supply taking the indexes into account, so that the price will depend on the effective contribution of each generator to the voltage rise in the target area. Simulations are presented to illustrate that the generator choice depends upon its localization. The need of reactive power support to improve the security margin against voltage collapse can be indicated by several indexes. It is shown a top view of the voltage collapse proximity indicators, with emphasis to those selected for this work. Often the electric power system operators call for a reactive power redispatch for voltage profile correction, but it is shown that sometimes this procedure leads to a risk of voltage collapse The dynamical behavior of synchronous generator excitation systems is very important for the reactive power dispatch. The operator requirements from Brazil, USA and Canada are presented. It is proposed a technique for voltage regulator adjustment, taking into account the time constant, the regulation error and the ceiling voltage, with illustration by means of simulations. Concluding, several suggestions are synthesized towards the evolution of the legislation concerning to reactive power supply from synchronous generators.

This work aimed to assess the relevance of the most common clinical tests and risk factors current used to establish the probability of falls among the elderly. 85 seniors took part in this research, both male and female, fallers and no fallers, 60 years old or more. This population was analyzed and divided into seven classes: (1) The whole population; (2) The whole female population (Female Group FG); (3) The whole male population (Male Group MG); (4) Female group 1 (FG1: 60 to 69 years old); (5) Female group 2 (FG2: > 70 years old); (6) Male group 1 (MG1: 60 to 69 years old); (7) Male group 2 (MG2: > 70 years old). The following criteria were considered in the analyses: quantity of falls, daily medication (Med), Body Mass Index (BMI), Calf Circumference (CC), Waist Circumference (WC), grip strength (MF), Dynamic Gait Index (DGI), balance (Berg Balance Scale BBS), mobility (Timed Up & Go TUG), cognitive state (Mini Mental State Examination MMSE), Life Satisfaction (LS), Physical Self-Efficacy (PSE), Functional Capacity (FC) and Fear of Falling (Falls Efficacy Scale International - FES I Brasil). Although 60% of the whole population did not fall, FG falls more frequently (53%) than MG (14%). The age range analysis shows that FG2 falls the most (FG2: 62%; FG1: 39%; MG2: 27%; and MG1: 6%). The Spearman Test (p &#8804; 0,05) revealed that the relationship fall x criteria (significantly correlated) varies according to the group, as follows: FG: Med, WC, MF, DGI, FES I Brasil, TUG, LS, FC; FG1: there was no relevant correlation; FG2: DGI; MG: BBS and TUG; MG2: FES I Brasil (p<0,001). The covariance was assessed for all indices in relation to the variable fall. According to those analyses, the most relevant elements are: FG: Med, DGI, and FES I Brasil; MG: MF, DGI, BBS, and TUG; FG1: BMI, WC, FES I Brasil, and TUG; FG2: Med, CC, and DGI; MG1: MF, DGI, and TUG; MG2: BBS and FES I Brasil. The results shows that the proper risk factors and the assessment tools to be used to evaluate the probability of fall among the elderly are significantly different for males and females, as well as among the same sex but at different age range. It is therefore suggested that the assessment protocols should be developed taking these differences into consideration, in order to ensure better efficacy, better directions for the treatment and earlier interventions. Such specificity would contribute to reduce the amount of falls and, consequently, improve the quality of life of the elderly population.

This thesis presents a study of architecture for virtual environment distribution as support tool for learning multidisciplinary projects. Different distribution platforms have been evaluated with the objective to identify the one that with more efficiency allowed exactly that occurred alterations in an environment modify the behavior of others, that these are related to other areas of the knowledge. Building prototypes under the platform chosen for the distribution, following same methodology (modifying aspects of the data model) and having the latency, scalability and extensibility as method of comparison had demonstrated to which the best choice for building multidisciplinary virtual environment. Each prototype was constructed on the basis of distribution algorithms that had allowed the system to function correctly in possible situations of errors. Virtual environments of Biology (landscape with plants, water, light and land) and Chemistry (membrane and molecules) had been used having the phenomenon of the photosynthesis as study of case and relation between two environments. The system was evaluated by professors and students and the reached results had allowed concluding that it is efficient and applicable.

A great number of researchers believe that problem resolution is inherent from the great part of the teaching and learning processes. It is considered that this methodology presents a better effectiveness when applied cooperatively, in other words, an apprentice improves your acting in solving problems when accomplishes his activity in cooperation with other apprentices. In the context of computational systems applied to learning, Virtual Reality (VR) is emerging as a technology that motivates the apprentice, since it allows the navigation, visualization, manipulation and interaction with virtual objects of its learning environment. In this context, this work presents a proposal that allows the use of VR to offers support to the resolution of problems through the cooperative learning. For this, the learning foundations will be considered by experimentation and the effectiveness of theories and constructivities pedagogic principles. As case study it was chosen for developing the project in the Physics ambit. The present work approaches and bases its proposal through the state of the art description of the research in this area. It makes an analysis, supplies specifications of the project, introduces details of system developing, discusses that evaluation and finally, presents the conclusions and proposals future work for this project.

Fundação de Amparo a Pesquisa do Estado de Minas Gerais

As the Internet has become an enormous interconnected network, the information security today is very important to guarantee confidentiality, integrity and availability of computing resources. Advanced Intrusion Detections Systems (IDS) should be capable of identifying malicious actions that may compromise these guarantees, as quickly as possible. In this work, we present a hybrid approach for the IDS, with two different techniques: wavelets and artificial neural network. The wavelet is used to indicate to detect anomalous behavior on the network, while the neural networks are used to classify the attacks. A prototype was developed and evaluated data from simulation tests, on laboratory network and the KDD99 database. Besides the analysis of results of the other approaches, was also carried out a comparison with the learning vector quantization. Good results were obtained in all experiments, this demonstrating that the approach is very promising.

This research developed a method of evaluation for computer tools for education (FCE), called Structure Cognitive for Education Technologies Constructivists (ECoTEC). This methodology detects the construction of structures of logical-mathematical thinking through the proposed activities on FCE. In the ECoTEC conception was used a strategy of analysis which includes: organization of cognitive structure of logical-mathematical thinking according to Piaget operative theory; the differentiation of essential concepts of this theory; and the deconstruction of Piaget operative theory. It was made an ascendant unfolding of principal concepts and organized them in a sort of integrated network and additional. The structure allows comprehending the structures of thinking in the concrete operative periods and formal operative, according to the operative activities, the interaction and the cooperation. From this structure it was established a process of measurement. It was used as a strategy for development the paradigm GQM (Goal/Question/Metric), a technique for identification of significant metrics. To automate the collection for information to the user and assist in calculation of the metrics for the characterization of FCE was developed an instrument of optimization called Web-ECoTEC. The validation of Web-ECoTEC while methodology of evaluation of FCE revealed his potentiality to characterize these tools according to the structure of thinking developed. The evaluations generated by the methodology can guide teachers, educators and educational professionals in the selection of FCE, since that these evaluations showed the structures of logical-mathematical thinking which can be used with the FCE evaluated. The utilization of Web-ECoTEC for validation, also presented efficient in the collection of answers and in generation of results of characterization of FCE.

A Domain Specific Language is a specification language dedicated to a particular domain, representation technique, or solution searching method. On the other hand, a general-purpose programming language is a language designed with the goal of emulating Lambda Calculus or Turing Machine. Since general-purpose languages must accept any algorithm that can be executed by a Turing Machine, they do not contain a knowledge base of a specific domain, which makes them difficult to master by a professional who is not a specialist in Computer Science. Many fields of science and technology have well advanced domain specific languages: LATEX and XML for text processing, SQL for data base management, Matlab for engineering, etc. Designers of neural network applications do not have good domain specific languages. The main reason for this situation is that they are considered computer scientists, and supposed to know general-purpose languages and even low level languages close to Turing machines (languages for microprogramming). Since we know that this has no base on reality, the goal of this work is to develop methods of creating domain specific languages rooted on Lisp macros and on functional languages (Clean and Haskell) for prototypes. We believe that Lisp is easy enough to be mastered even by people who has difficulty with formal methods and mathematics. Therefore, with well designed extensions, and a rapid training in programming methods, an engineer can use Lisp as a powerful tool of productivity. Lisp is a traditional tool for creating both embedded languages and domain specific languages. Therefore, Lisp offers all tools necessary to design, test and deploy domain specific languages. Nevertheless, neural networks have features that require special care. For instance, they need to have biological feasibility. In order to reach code quality close to human designed nets, while preserving biological feasibility, this work will use advanced Artificial Intelligence methods and guided automatic theorem provers similar to the one used in HOL4 or Isabelle. Our system has a simpler goal than general theorem provers, since it will focus in a few problems related to neural networks. One of these problems is the designing of devices for data acquisition. This thesis differs from similar works in one point that may cause controversy: Besides the contributions of the author to the state of the art, it discusses elementary aspects of the technologies involved. We decided to do this in order to bring the ideas that we are exposing to the reach of all readers. What are these ideas? The first contribution is to bring software engineering technology to biological modeling. We are not aware of any researcher who attempted this goal before. A second aspect of this work can be resumed as an effort to create a language to describe combinations of neuron models into complex networks. All programs discussed here were designed to avoid obsolescence. The author achieved this goal by writing prototypes in ¸-calculus, a formal system that can be executed in modern computers.

This work consists of the study, design and implementation of DC-DC converters, used in lighting systems based on light emitting diodes (LEDs). Can be powered by an universal AC voltage source or battery, and the current in the LEDs is controlled digitally. The main objective is to present a proposal of new step-up/down converters topologies with full conversion much higher than that of conventional converters. So, is initially presented an overview of solid-state lighting (SSL). Following is presented the study on the dynamic behavior of converters based on modeling by the method of varying state-space average, ending with the design of a digital PI controller. This work resulted in the application for a patent PI-0801425-6 filed with the INPI, which presents a new family of converters with full conversion extremely wide, called cubic converters.

Universidade Federal do Triângulo Mineiro

There are several techniques applied to optimization problems. However, few achieve satisfactory performance when the problem is complex, for example multi-modal or multi- objective. The metaheuristics, examples of optimization techniques, are heuristic algo- rithms that can not guarantee the global optimum, but usually ¯nd good solutions. There are several metaheuristics, as an example: Simulated Anneling, Tabu Search, GRASP, VND, VNS and Ant Colony. Among the metaheuristics, the algorithms of Evolu- tionary Computation is one of the most used, since the e®ectiveness and modular/adaptive features. Evolution strategy, Genetic programming and Evolutionary programming are examples of evolutionary algorithms, however, the pioneer and the most popular in the literature is the Genetic Algorithm, despite the di±culties of convergence in some cases. Genetic Algorithms (GA) are optimization and search methods inspired by the evolu- tion of living beings population. Algorithms, based on this technique, follow the principle of natural selection and survival of the fittest (Charles Darwin). When we analyze the GA, where several generations are produced, one on each itera- tion, and considering that each new generation the old one is partially or totally discarded, the GA may be removing relevant information within individuals discarded, which were not transmitted or evaluated by the algorithm. Therefore, and considering the few studies that address the improvement in the use of the information, and the need to present evolutionary solutions with wider applicability, this paper proposes the Algoritmo Auxiliar Paralelo (AAP), which aims to assist the GA with good individuals from better treatment of the structures present in parents populations. The AAP is an auxiliary module running on a parallel °ow to the GA and that recombine chromosomes to maximize the use of the information present in individuals. As a result, the module can generate artificial fittest individuals, which are inserted in the new generation and manipulated by the GA in the next iteration. Inspired by In Vitro Fertilization and the Preimplantation Genetic Diagnosis, which reviews and selects good pre-embryos to be transferred to the mother, the AAP following a °ow of Collection, Manipulate, Select and Transfer of good individuals. To test the performance of the proposed algorithm (AAP), and their operators, when linked to the GA, we chose two benchmark problems. The ¯rst, Rastrigin function by ha- ving a large search space, non-linear and have a high degree of multimodality. The second, the Multidimensional Knapsack Problem, as a multidimensional and highly combinatorial problem. Therefore, it was possible to measure the performance of the proposal in two dif- ferent types of problem: non-restrictive multimodal function and restrictive combinatorial optimization. The AAP has been tested and compared with the canonical GA, identifying the performance of the proposal, and with a hybrid algorithm GA-TS (Tabu Search), which has characteristics of global and local search. We constructed 39 sets of the addressed problems for the tests. The results show the AAP as a good tool to assist the GA to improve the convergence performance. It is noticed also that there was a considerable improvement in the speed of convergence without a®ecting the quality of the ¯nal solution. Given the results and the modular structure of AAP, allowing other changes and new operators, we believed that the AAP may be useful in various applications and applicable to other populations heuristics.

In recent years, there is a growing interest in many-valued logics in many areas such as computer sciences automated theorem proving, approximate reasoning, multi-agent systems, program verication electrical engineering and digital circuits, linguistics, mathematics and algebra, philosophy, etc. In this context, the general problem of nding an axiomatization for nite many-valued logics has not yet been solved satisfactorily. For this, we must demonstrate the correction and completeness theorems of these logic systems, both mathematically rigorous. Soundness theorem is straightforward, however, it is known that the demonstration of the "completeness theorem"is much more sosticated and unique for each logic system. Completeness is one of the most important notions in logic and the foundations of mathematics. Completeness means the possibility of getting all correct and reliable schemata of inference by use of logical methods. When one wishes to build or design a many-valued logic system, he or she is implicitly looking for a system with a minimal set of axioms and rules. The purpose of this thesis is to establish a general algorithmic completeness proof procedure for nite many-valued logics. It is shown that a matrix is characteristic (sound and complete) for a many-valued nite logic system when, all successive correct extensions of this matrix have the same set of tautologies. It is also shown that in order to determine if a matrix is characteristic, all one has to do is to fetch from all the unit extensions those which are correct and then verify if they are repetitions of the matrix being considered. How to implement a computational procedure to demonstrate the completeness of a many-valued nite logic system is shown from these results. The new approach is a simpler and more uniform solution for the problem of nite many-valued logics axiomatization.

The Teaching and learning computer-programming process as been the target of scientific research since the early days of computing. Although several strategies have been and are still being tested, its application in computer-based courses has been difficult, since it requires some commitment from educators to search, study and evaluate them. Although not looking for a miraculous cure for all problems in this area (reported in the literature), and believing in the pedagogical potential of the socio-cognitive conflict theory, it has been decided to develop a teaching method of programming based on Bloom's taxonomy of educational objectives to be applied using the pair-programming technique. This method is being used to produce a book of Algorithms and Programming Techniques intending to be acknowledged by the Academy. Experiments with such method has been carried out at a private University and there has been a significant increase of motivation by the students that underwent the experiments.

Fundação de Amparo a Pesquisa do Estado de Minas Gerais

Fundação de Amparo a Pesquisa do Estado de Minas Gerais

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior

The Internet and ebooks are replacing traditional publishing houses. Encyclopedias, for instance, are published mainly in the Internet, since this media allows fast update, cooperation and corrections. Electronic readers decreased the cost of publishing books, and spare natural resources (wood for papers, for instance) needed for printting documents Since long, publishing technology have advanced two domain specific languages for describing documents: LATEX and XML. In recent years, XML became the choice tool for the Internet, text processing, and electronic edition of books. Therefore, this dissertation will offer tools for producing XML for www pages, and also for electronic publishing. The Scheme programming language is a dialect of LISP well suited for text processing, since it has a good representation for structured text: Nested lists. Thus the author of this work will describe how to use Scheme for creating dynamic pages based on their XML description. If global communication is to become feasible, we must adapt the technological advances that people who work with Artificial Intelligence made in the fields of knowledge representation, context capture, domain specific languages, reflexivity, and natural language processing. In general, AI technologies are expressed in languages of the LISP family, since they are homoiconic, i.e., their primary representation of programs is also a data structure in a primitive type of LISP. This makes metaprogramming easier than in a language without this property. This paper will show how to use metaprogramming and closures for skimming through a restricted domain text in order to grasp its syntactic structure and meaning. Once the software system obtains the general features of a text, it can process it to generate books that can be installed in dedicated reading devices; it can also generate dictionaries and digests to easy the process of reading. Finally this work will address the problem of publishing ebooks, after specifying their structure in XML.