Roberta Vilhena Vieira Lopes

An Application of Genetic Algorithm to the Game of Checkers

The difficulty of a games Artificial Intelligence is one of the determining factors of its success or failure. To define how the machine-opponent will behave is a hard and complex task. This paper aims to present an application of the Game of Checkers that tries to solve the matter of presenting an adversary which is convincing and of a similar level to the player, using Genetic Algorithms techniques.

Um Detector de Complexos QRS Evolutivo para o Eletrocardiograma

Resumo — Este trabalho descreve a implementação e utilização de um algoritmo genético que possui como finalidade a detecção dos complexos QRS de um Eletrocardiograma. Os dados obtidos com a execução do algoritmo são importantes para a determinação da morfologia dos complexos QRS, além de ser possível utilizar estas informações para a obtenção de dados como a freqüência cardíaca e a presença de arritmias. O desempenho do algoritmo foi avaliado através de testes realizados com amostras da base de dados MIT-BIH Arrhythmia Database e o resultado desta que atestam a viabilidade do emprego do algoritmo.

An application of algorithm based on abstract data types to the keratoconus diagnosis

In this paper, it is proposed a new genetic algorithm based on abstract data types approach for determination of coefficients can be adjusted by evolutionary process considering some metrics classification for keratoconus patients parameters from the ORA®. It is intended to show the efficiency of genetic algorithm based on abstract data types in the characterization of a combined solution for complex optimization problems, like the classifier.

GAADT (Genetic Algorithm based on Abstracts Data Types): Uma Abordagem Teórica e Aplicativa

This article presents an overview of the Theory of Genetic Algorithms based on Abstract Data Types. Here are some epistemological questions raised about the use of a logic deduced from the Theory of the Origin of Species Darwin (Darwin, 1959), (Gould, 1992) to justify the change in level of abstraction as a vision in search of an appropriate methodology for solving certain types of problems that are here called Evolutionary problems (Vieira, 2003). For this reason, this theory deals with a representation of the individual epistemological stratified into three levels of perception: chromosome, gene and base and the population as a collection of individuals based on these inputs. This perception is sufficiently rich enough to allow evolution to be seen as a move in a certain direction determined by the environment. The logical result of this view of evolution is used here to solve problems. With this theory it is possible to characterize certain properties of genetic algorithms, to identify problems that can be solved by them, and make comparisons with other models. The adoption of a rigorous language for the description of genetic algorithms and, consequently, the establishment of a formal theory for the treatment of the same, give the stamp of scientific and a set of applicability criteria that currently lack these powerful mechanisms. Finally, a method is presented, and the application that is developed during the work.

A MMORPG decision-making model based on persuasive reciprocity

From a videogame perspective, decision-making is a crucial activity that happens at all times and at different levels of perception. Moreover, it has direct influence over a games performance, a fact that also concerns to RPGs as they can act as tools to enhance the improvement of the proximal development zones of the involved individuals. As the RPG has an inherent cooperative character that stimulates socialization, interaction and the improvement of communication skills, it was thought that it would be interesting to take advantage of the RPG to build a model using Petri Net, based on the Game Theory and on a application of the Theory of Persuasion to an MMORPG environment, that involves the user in some kind of plot, at the same time that favors a greater ease in decision-making.

A variation of the genetic algorithm of Holland to support analysis of balance sheet and Income Statement for the Fiscal Year

Accounting is a science that, by studying the financial statements, search to provide information of great importance for decision-making, becoming a crucial tool for the managers of an organization. One of the existing difficulties is in the transparency and understanding of accounting data by managers. The studies in this area show the use of subjectivity on the part of financial analysts while in analysis of financial statements. This is due to the fact that the analysis involves deterministic and nondeterministic factors, such as influence of the branch of business, the companys operating cycle, seasons, political influence, etc. Therefore, this article proposes the creation of a evolutionary application to be used as a tool to support the analysis of accounting indicators, providing a more accurate view of the result by reducing the subjectivity inherent to the analyst. The application of this tool used data extracted from the Balance Sheets (BS) and Income Statement for the Fiscal Year (SFY) of publicly traded corporations Arcelor Mittal Brazil S/A and Romi Industries S.A. for analysis of indicators.

An application of a new genetic algorithm to assist the detection of segments of an ECG signals

In this paper, we present the specification of a type of genetic algorithm that was used for the detection of segments present in ECG signal waves. The algorithm was tested and evaluated for its efficiency in front of samples obtained from the MIT-BIH Arrhythmia Database and the results are also presented here. The algorithm used was based on a genetic algorithm based on abstract data types and was implemented in MatLab. The results showed that the detector can be implemented in systems for detecting arrhythmias and may become a tool to improve medical care treatment that is provided to patients with cardiovascular problems.

Application of a Genetic Algorithm Based on Abstract Data Type in Electronic Games' Scenarios Adaptive Evolution

The importance of artificial intelligence in electronic games has increased in response to the requirements of a more realistic behavior. For the application of such techniques, an excellent candidate is the RPG game genre which provides rich and interactive environments that, similarly to a real world environment, have interesting events taking place simultaneously. Using a RPG game as background, this paper aims to demonstrate the application of a genetic algorithm, specifically a genetic algorithm based on abstract data types (GAADT), to optimize the build of game scenarios, using an evaluation function that seeks and inserts a new characteristic into a game map employing its best efforts to make the transition to such characteristic as gradual as possible, giving as result a cohesive map that has its geographic characteristics as an extra factor to be considered by the players society.