XIII Congreso Argentino de Ciencias de la Computación http://sedici.unlp.edu.ar:80/handle/10915/21589 2026-04-15T13:37:12Z 2026-04-15T13:37:12Z Cecchini, Rocío L. Maguitman, Ana Gabriela Lorenzetti, Carlos M. http://sedici.unlp.edu.ar:80/handle/10915/23581 2018-12-23T04:04:20Z 2007-01-01T00:00:00Z

Objeto de conferencia XIII Congreso Argentino de Ciencias de la Computación Successful access to information sources on the Web depends on effective methods for identifying the needs of a user and making relevant information resources available when needed. This paper formulates a theoretical framework for the study of context-drivenWeb search and proposes new methods for learning query terms based on the user task. These methods use an incrementally-retrieved, topic-dependent selection of Web documents for term-weight reinforcement reflecting the aptness of the terms in describing and discriminating the topic of the user context. Based on this framework, we propose an incremental search algorithm for information retrieval agents that has the potential to improve significantly over the traditional IR techniques. The new algorithm learns new descriptors by searching for terms that tend to occur often in relevant documents, and learns good discriminators by identifying terms that tend to occur only in the context of the given topic. We discuss the technical challenges posed by this new framework, outline our agent system architecture, and present an evaluation of the proposed techniques.

2007-01-01T00:00:00Z Successful access to information sources on the Web depends on effective methods for identifying the needs of a user and making relevant information resources available when needed. This paper formulates a theoretical framework for the study of context-drivenWeb search and proposes new methods for learning query terms based on the user task. These methods use an incrementally-retrieved, topic-dependent selection of Web documents for term-weight reinforcement reflecting the aptness of the terms in describing and discriminating the topic of the user context. Based on this framework, we propose an incremental search algorithm for information retrieval agents that has the potential to improve significantly over the traditional IR techniques. The new algorithm learns new descriptors by searching for terms that tend to occur often in relevant documents, and learns good discriminators by identifying terms that tend to occur only in the context of the given topic. We discuss the technical challenges posed by this new framework, outline our agent system architecture, and present an evaluation of the proposed techniques. Sagui, Fernando M. Chesñevar, Carlos Iván Maguitman, Ana Gabriela Simari, Guillermo Ricardo http://sedici.unlp.edu.ar:80/handle/10915/23580 2018-12-23T04:04:44Z 2007-01-01T00:00:00Z

Objeto de conferencia XIII Congreso Argentino de Ciencias de la Computación Due to the huge amount of multi-source news that are available on the Web at any time, it is crucial to provide intelligent mechanisms to select and rank news reports. Over the last few years, a number of approaches based on criteria such as freshness, relevance and viewer profile have been proposed. However, most existing news processing services do not deal with credibility from a qualitative perspective, and do not provide mechanisms to cope with controversial news reports. To fill this gap, this paper proposes a news service framework that brings the notions of trust and pluralism into play. The proposed framework is based on a set of basic postulates characterizing the nature of trust. In our proposal, trust is modeled using Defeasible Logic Programming, a general-purpose defeasible argumentation formalism based on logic programming. Our approach helps identify antagonism among sources of news and facilitates the analysis of opposing positions. This allows us to integrate dialectical reasoning into a news recommender system, which has the capability of providing a reasoned basis for the news presented to the viewer.

2007-01-01T00:00:00Z Due to the huge amount of multi-source news that are available on the Web at any time, it is crucial to provide intelligent mechanisms to select and rank news reports. Over the last few years, a number of approaches based on criteria such as freshness, relevance and viewer profile have been proposed. However, most existing news processing services do not deal with credibility from a qualitative perspective, and do not provide mechanisms to cope with controversial news reports. To fill this gap, this paper proposes a news service framework that brings the notions of trust and pluralism into play. The proposed framework is based on a set of basic postulates characterizing the nature of trust. In our proposal, trust is modeled using Defeasible Logic Programming, a general-purpose defeasible argumentation formalism based on logic programming. Our approach helps identify antagonism among sources of news and facilitates the analysis of opposing positions. This allows us to integrate dialectical reasoning into a news recommender system, which has the capability of providing a reasoned basis for the news presented to the viewer. Cecchini, Rocío L. Lorenzetti, Carlos M. Maguitman, Ana Gabriela Brignole, Nélida B. http://sedici.unlp.edu.ar:80/handle/10915/23579 2018-12-23T04:04:38Z 2007-01-01T00:00:00Z

Objeto de conferencia XIII Congreso Argentino de Ciencias de la Computación Harvesting topical content is a process that can be done by formulating topic-relevant queries and submitting them to a search engine. The quality of the material collected through this process is highly dependant on the vocabulary used to generate the search queries. In this scenario, selecting good query terms can be seen as an optimization problem where the objective function to be optimized is based on the effectiveness of a query to retrieve relevant material. Three characteristics of this optimization problem are (1) the high-dimensionality of the search space, where candidate solutions are queries and each term corresponds to a different dimension, (2) the existence of acceptable suboptimal solutions, and (3) the possibility of finding multiple solutions. This paper describes optimization techniques based on Genetic Algorithms to evolve “good query terms” in the context of a given topic. We discuss the use of a mutation pool to allow the generation of queries with novel terms, and study the effect of different mutation rates on the exploration of query-space.

2007-01-01T00:00:00Z Harvesting topical content is a process that can be done by formulating topic-relevant queries and submitting them to a search engine. The quality of the material collected through this process is highly dependant on the vocabulary used to generate the search queries. In this scenario, selecting good query terms can be seen as an optimization problem where the objective function to be optimized is based on the effectiveness of a query to retrieve relevant material. Three characteristics of this optimization problem are (1) the high-dimensionality of the search space, where candidate solutions are queries and each term corresponds to a different dimension, (2) the existence of acceptable suboptimal solutions, and (3) the possibility of finding multiple solutions. This paper describes optimization techniques based on Genetic Algorithms to evolve “good query terms” in the context of a given topic. We discuss the use of a mutation pool to allow the generation of queries with novel terms, and study the effect of different mutation rates on the exploration of query-space. Gómez Lucero, Mauro Javier García, Alejandro Javier Chesñevar, Carlos Iván Simari, Guillermo Ricardo http://sedici.unlp.edu.ar:80/handle/10915/23578 2018-12-23T04:05:02Z 2007-01-01T00:00:00Z

Objeto de conferencia XIII Congreso Argentino de Ciencias de la Computación This article presents the confluence of two general ideas: Defeasible Logic Programming (DeLP, an argumentation based formalism for representing knowledge and reasoning) and Abduction in logic. In first place, we introduce a framework that formally states the problem of doing abduction (obtaining abductive explanations) in the particular case that the abductive theory is a DeLP Program. In the reminder of this work we face that problem, providing a formal characterization of the notion of abductive explanation, in such a way that we could then easily calculate the explanations from this formal characterization. An important advantage of this approach is that by proving the correctness of the characterization, we ensure the correctness of the method for obtaining explanations.

2007-01-01T00:00:00Z This article presents the confluence of two general ideas: Defeasible Logic Programming (DeLP, an argumentation based formalism for representing knowledge and reasoning) and Abduction in logic. In first place, we introduce a framework that formally states the problem of doing abduction (obtaining abductive explanations) in the particular case that the abductive theory is a DeLP Program. In the reminder of this work we face that problem, providing a formal characterization of the notion of abductive explanation, in such a way that we could then easily calculate the explanations from this formal characterization. An important advantage of this approach is that by proving the correctness of the characterization, we ensure the correctness of the method for obtaining explanations. Tucat, Mariano García, Alejandro Javier http://sedici.unlp.edu.ar:80/handle/10915/23577 2018-12-23T04:04:57Z 2007-01-01T00:00:00Z

Objeto de conferencia XIII Congreso Argentino de Ciencias de la Computación In this paper we consider the design of an agent communication model based on the FIPA Architecture and FIPA Interaction Protocols. Our goal is to allow the agents to reach each other by their characteristics and to exchange messages using a standard Agent Communication Language. We propose the design and implementation of a framework as an extension of Prolog, following the spirit of Logic Programming. In our framework, the agents are allowed to register themselves in MASs in order to allow other agents to locate them easily. The agents may search for agents having specific characteristics, or providing determined services in order to interact with them.

2007-01-01T00:00:00Z In this paper we consider the design of an agent communication model based on the FIPA Architecture and FIPA Interaction Protocols. Our goal is to allow the agents to reach each other by their characteristics and to exchange messages using a standard Agent Communication Language. We propose the design and implementation of a framework as an extension of Prolog, following the spirit of Logic Programming. In our framework, the agents are allowed to register themselves in MASs in order to allow other agents to locate them easily. The agents may search for agents having specific characteristics, or providing determined services in order to interact with them. Salto, Carolina Alba Torres, Enrique Molina, J.M. Leguizamón, Guillermo http://sedici.unlp.edu.ar:80/handle/10915/23576 2018-12-23T04:05:30Z 2007-01-01T00:00:00Z

Objeto de conferencia XIII Congreso Argentino de Ciencias de la Computación In this paper, the two-dimensional strip packing problem with 3-stage level patterns is tackled using genetic algorithms (GAs). We evaluate the usefulness of a greedy seeding procedure for creating the initial population, incorporating problem knowledge. This is motivated by the expectation that the seeding will speed up the GA by starting the search in promising regions of the search space. An analysis of the impact of the seeded initial population is offered, together with a complete study of the influence of these modifications on the genetic search. The results show that the use of an appropriate seeding of the initial population outperforms existing GA approaches on all the used problem instances, for all the metrics used, and in fact it represents the new state of the art for this problem.

2007-01-01T00:00:00Z In this paper, the two-dimensional strip packing problem with 3-stage level patterns is tackled using genetic algorithms (GAs). We evaluate the usefulness of a greedy seeding procedure for creating the initial population, incorporating problem knowledge. This is motivated by the expectation that the seeding will speed up the GA by starting the search in promising regions of the search space. An analysis of the impact of the seeded initial population is offered, together with a complete study of the influence of these modifications on the genetic search. The results show that the use of an appropriate seeding of the initial population outperforms existing GA approaches on all the used problem instances, for all the metrics used, and in fact it represents the new state of the art for this problem. Casali, Ana Von Furth, Armando Sierra, Carles Godo, Lluís http://sedici.unlp.edu.ar:80/handle/10915/23575 2018-12-23T04:04:51Z 2007-01-01T00:00:00Z

Objeto de conferencia XIII Congreso Argentino de Ciencias de la Computación In this paper a multiagent Tourism Recommender System is presented. This system has a multiagent architecture and one of its main agents, The Travel Assistant Agent (T-Agent), is modelled as a graded BDI agent. The graded BDI agent model allows to specify an agent’s architecture able to deal with the environment uncertainty and with graded mental attitudes. We focus on the implementational aspects of the multiagent system and specially on the T-Agent development, going from the theoric agent model to the concrete agent implementation.

2007-01-01T00:00:00Z In this paper a multiagent Tourism Recommender System is presented. This system has a multiagent architecture and one of its main agents, The Travel Assistant Agent (T-Agent), is modelled as a graded BDI agent. The graded BDI agent model allows to specify an agent’s architecture able to deal with the environment uncertainty and with graded mental attitudes. We focus on the implementational aspects of the multiagent system and specially on the T-Agent development, going from the theoric agent model to the concrete agent implementation. Falappa, Marcelo Alejandro Moguillansky, Martín O. Rotstein, Nicolás D. Simari, Guillermo Ricardo http://sedici.unlp.edu.ar:80/handle/10915/23574 2018-12-23T04:05:24Z 2007-01-01T00:00:00Z

Objeto de conferencia XIII Congreso Argentino de Ciencias de la Computación In this article we are presenting a new perspective on the matter of belief revision by its relation to argumentation systems. Our approach is based on the argumentative formalism Defeasible Logic Programming, and therefore we propose a revision of a defeasible logic program by an argument. The revision operators here introduced are de ned as prioritized, since they ensure warrant of the conclusion of the argument being added to the program following a particular minimal change principle. To achieve this, we give two different approaches: one regarding arguments in the classical sense, whereas the other considers the revision by arguments that also include strict rules and facts. Finally, a brief discussion about the relation between our approach and the basic theory of belief revision is exposed, along with a description of other possible minimal change principles.

2007-01-01T00:00:00Z In this article we are presenting a new perspective on the matter of belief revision by its relation to argumentation systems. Our approach is based on the argumentative formalism Defeasible Logic Programming, and therefore we propose a revision of a defeasible logic program by an argument. The revision operators here introduced are de ned as prioritized, since they ensure warrant of the conclusion of the argument being added to the program following a particular minimal change principle. To achieve this, we give two different approaches: one regarding arguments in the classical sense, whereas the other considers the revision by arguments that also include strict rules and facts. Finally, a brief discussion about the relation between our approach and the basic theory of belief revision is exposed, along with a description of other possible minimal change principles. Falappa, Marcelo Alejandro Moguillansky, Martín O. http://sedici.unlp.edu.ar:80/handle/10915/23573 2018-12-23T04:05:19Z 2007-01-01T00:00:00Z

Objeto de conferencia XIII Congreso Argentino de Ciencias de la Computación When using tableaux algorithms to reason about Description Logics (DLs), new information is inferred from the models considered while trying to achieve knowledge satisfiability. By focusing the ontology change problem, we consider an environment where DLs are the logical formalization to express knowledge bases in the web, and the integration of distributed ontologies is developed under new extensions of the belief revision theories originally exposed in [1]. Hence, a reinforced theory arises in order to properly apply change operations over models, considering new inferred information and assumed beliefs in each possible world. As a result, a new type of contraction operator is proposed and its success postulate analyzed.

2007-01-01T00:00:00Z When using tableaux algorithms to reason about Description Logics (DLs), new information is inferred from the models considered while trying to achieve knowledge satisfiability. By focusing the ontology change problem, we consider an environment where DLs are the logical formalization to express knowledge bases in the web, and the integration of distributed ontologies is developed under new extensions of the belief revision theories originally exposed in [1]. Hence, a reinforced theory arises in order to properly apply change operations over models, considering new inferred information and assumed beliefs in each possible world. As a result, a new type of contraction operator is proposed and its success postulate analyzed. Martínez, Diego C. García, Alejandro Javier Simari, Guillermo Ricardo http://sedici.unlp.edu.ar:80/handle/10915/23486 2018-12-23T04:04:32Z 2007-01-01T00:00:00Z

Objeto de conferencia XIII Congreso Argentino de Ciencias de la Computación In this work, contexts for extended argumentation frameworks (EAF) are defined. A context for an EAF is another framework where original arguments, conflicts and preferences are kept, while introducing new arguments leading to new defeat relations. Thus, the context may interfere with the original classification of arguments, inducing new set of extensions. These semantic change in the outcome of an extended framework in a particular context is characterized, and Dung’s acceptability concept is analyzed on this basis.

2007-01-01T00:00:00Z In this work, contexts for extended argumentation frameworks (EAF) are defined. A context for an EAF is another framework where original arguments, conflicts and preferences are kept, while introducing new arguments leading to new defeat relations. Thus, the context may interfere with the original classification of arguments, inducing new set of extensions. These semantic change in the outcome of an extended framework in a particular context is characterized, and Dung’s acceptability concept is analyzed on this basis. Namías, Rafael Granitto, Pablo Miguel http://sedici.unlp.edu.ar:80/handle/10915/23485 2018-12-23T04:05:14Z 2007-01-01T00:00:00Z

Objeto de conferencia XIII Congreso Argentino de Ciencias de la Computación Ensemble methods show improved generalization capabilities that outperforrn those of single larners. lt is generally accepted that, for aggregation to be effective, the individual learners must be as accurate and diverse as possible. An important problem in ensemble learning is then how to find a good balance between these two conflicting conditions. For tree-based methods a successfill strategy was introduced by Breiman with the Random-Forest algorithm. In this work we introduce new methods for neural network ensemble construction that follow Random-Forest-like strategies to construct ensembles. Using several real and artificial regression problems, we compare onr new methods with the more typical Bagging algorithrm and with three state-of-the-art regression methods. We find that our algorithms produce very good results on several datasets. Some evidence suggest that our new methods work better on problems with several redundant or noisy inputs.

2007-01-01T00:00:00Z Ensemble methods show improved generalization capabilities that outperforrn those of single larners. lt is generally accepted that, for aggregation to be effective, the individual learners must be as accurate and diverse as possible. An important problem in ensemble learning is then how to find a good balance between these two conflicting conditions. For tree-based methods a successfill strategy was introduced by Breiman with the Random-Forest algorithm. In this work we introduce new methods for neural network ensemble construction that follow Random-Forest-like strategies to construct ensembles. Using several real and artificial regression problems, we compare onr new methods with the more typical Bagging algorithrm and with three state-of-the-art regression methods. We find that our algorithms produce very good results on several datasets. Some evidence suggest that our new methods work better on problems with several redundant or noisy inputs. Grinblat, Guillermo L. Ceccatto, Hermenegildo Alejandro http://sedici.unlp.edu.ar:80/handle/10915/23484 2018-12-23T04:05:08Z 2007-01-01T00:00:00Z

Objeto de conferencia XIII Congreso Argentino de Ciencias de la Computación Se propone un método de clasificación adaptativo capaz de aprender un concepto y seguir su evolución temporal como consecuencia de cambios lentos en sistemas evolutivos. Para ello se realiza una modificación del clasificador SVM (máquina de vectores soporte), consistente en usar múltiples hiperplanos válidos en pequeñas localidades temporales (ventanas) para realizar la clasificación. A diferencia de otras propuestas de este tipo en la literatura, en este caso se realiza un aprendizaje de todos los hiperplanos en forma global, minimizando una cantidad que contiene al error que comete la familia de clasificadores locales más una medida asociada a la dimensión VC de los mismos. Para conceptos estacionarios, la misma idea aplicada a localidades en el espacio de características permite obtener resultados comparables a los que proporciona SVM con kernel gausiano.

2007-01-01T00:00:00Z Se propone un método de clasificación adaptativo capaz de aprender un concepto y seguir su evolución temporal como consecuencia de cambios lentos en sistemas evolutivos. Para ello se realiza una modificación del clasificador SVM (máquina de vectores soporte), consistente en usar múltiples hiperplanos válidos en pequeñas localidades temporales (ventanas) para realizar la clasificación. A diferencia de otras propuestas de este tipo en la literatura, en este caso se realiza un aprendizaje de todos los hiperplanos en forma global, minimizando una cantidad que contiene al error que comete la familia de clasificadores locales más una medida asociada a la dimensión VC de los mismos. Para conceptos estacionarios, la misma idea aplicada a localidades en el espacio de características permite obtener resultados comparables a los que proporciona SVM con kernel gausiano. Gottifredi, Sebastián Simari, Guillermo Ricardo García, Alejandro Javier http://sedici.unlp.edu.ar:80/handle/10915/23483 2018-12-23T04:04:26Z 2007-01-01T00:00:00Z

Objeto de conferencia XIII Congreso Argentino de Ciencias de la Computación In this work two declarative approaches based on the BDI theory are studied, an agent programming language 3APL and an agent architecture that uses defeasible argumentation for knowledge representation and reasoning. Based on that study and considering that in 3APL the knowledge representation language is not fixed, we will propose 3APL-DeLP where the agent knowledge is represented by a DeLP-program and the agent may reason with a defeasible argumentation formalism.

2007-01-01T00:00:00Z In this work two declarative approaches based on the BDI theory are studied, an agent programming language 3APL and an agent architecture that uses defeasible argumentation for knowledge representation and reasoning. Based on that study and considering that in 3APL the knowledge representation language is not fixed, we will propose 3APL-DeLP where the agent knowledge is represented by a DeLP-program and the agent may reason with a defeasible argumentation formalism. Dolz, Daniel Parra, Gerardo http://sedici.unlp.edu.ar:80/handle/10915/23376 2018-12-22T20:05:12Z 2007-01-01T00:00:00Z

Objeto de conferencia XIII Congreso Argentino de Ciencias de la Computación La plataforma .NET de Microsoft se basa en una filosofía de just-in-time compilation (compilación bajo demanda al momento de la ejecución). Los programas desarrollados de esta manera se ejecutan en un entorno o framework independiente de la plataforma, basado en objetos y, en algunos casos, permitiendo que interactúen componentes desarrollados en distintos lenguajes de programación. La clave de esta flexibilidad se da en que, tanto en JAVA como en las plataformas .NET, la compilación resulta en un código intermedio, independiente de la plataforma (bytecode y MSIL respectivamente). Sin embargo, dicha flexibilidad tiene un costo. Hoy en día, y utilizando herramientas gratuitas que pueden descargarse desde Internet, es sumamente fácil aplicar tecnologías de ingeniería inversa a las dos plataformas de desarrollo más populares: JAVA y .NET. De todas las técnicas que los desarrolladores pueden utilizar para proteger su propiedad intelectual, la ofuscación es la técnica más aceptada y de hecho, es la única utilizada comercialmente. En este trabajo, proponemos el uso de los mecanismos de excepción (bloques try-catch) que brinda la plataforma .NET como una manera de mejorar la calidad de uno de los bloques básicos de la ofuscación, los predicados opacos.; Microsoft’s .NET Framework, and JAVA platforms, are based in a just-in-time compilation philosophy. Software developed using these technologies are executed in a hardware independent framework, which provides a full object-oriented environment, and in some cases allows the interaction of several components written in different programming languages. This flexibility is achieved by compiling into an intermediate code which is platform independent. Java is compiled into ByteCode, and Microsoft .NET programs are compiled into MSIL (Microsoft Intermediate Code). However, this flexibility comes with a price. It is really easy, with tools available for free in the web, to decompile this intermediate code and obtain a working, readable version of the original source code. Of all techniques developers can use to protect their intellectual property, obfuscation is the most accepted and commercially available one. In the present work, we propose the use of try-catch mechanisms available in .NET as a way to improve the quality of one of the building blocks of obfuscation: opaques predicates.

2007-01-01T00:00:00Z La plataforma .NET de Microsoft se basa en una filosofía de just-in-time compilation (compilación bajo demanda al momento de la ejecución). Los programas desarrollados de esta manera se ejecutan en un entorno o framework independiente de la plataforma, basado en objetos y, en algunos casos, permitiendo que interactúen componentes desarrollados en distintos lenguajes de programación. La clave de esta flexibilidad se da en que, tanto en JAVA como en las plataformas .NET, la compilación resulta en un código intermedio, independiente de la plataforma (bytecode y MSIL respectivamente). Sin embargo, dicha flexibilidad tiene un costo. Hoy en día, y utilizando herramientas gratuitas que pueden descargarse desde Internet, es sumamente fácil aplicar tecnologías de ingeniería inversa a las dos plataformas de desarrollo más populares: JAVA y .NET. De todas las técnicas que los desarrolladores pueden utilizar para proteger su propiedad intelectual, la ofuscación es la técnica más aceptada y de hecho, es la única utilizada comercialmente. En este trabajo, proponemos el uso de los mecanismos de excepción (bloques try-catch) que brinda la plataforma .NET como una manera de mejorar la calidad de uno de los bloques básicos de la ofuscación, los predicados opacos. Microsoft’s .NET Framework, and JAVA platforms, are based in a just-in-time compilation philosophy. Software developed using these technologies are executed in a hardware independent framework, which provides a full object-oriented environment, and in some cases allows the interaction of several components written in different programming languages. This flexibility is achieved by compiling into an intermediate code which is platform independent. Java is compiled into ByteCode, and Microsoft .NET programs are compiled into MSIL (Microsoft Intermediate Code). However, this flexibility comes with a price. It is really easy, with tools available for free in the web, to decompile this intermediate code and obtain a working, readable version of the original source code. Of all techniques developers can use to protect their intellectual property, obfuscation is the most accepted and commercially available one. In the present work, we propose the use of try-catch mechanisms available in .NET as a way to improve the quality of one of the building blocks of obfuscation: opaques predicates. Osorio, L. Fernández, Enrique Merlino, Hernán Rodríguez, Dario Britos, Paola Verónica García Martínez, Ramón http://sedici.unlp.edu.ar:80/handle/10915/23375 2018-12-23T04:02:41Z 2007-01-01T00:00:00Z

Objeto de conferencia XIII Congreso Argentino de Ciencias de la Computación Los algoritmos Genéticos, y su generalización: los Algoritmos Evolucionarios; presentan una alternativa para solucionar problemas de optimización complejos. Si bien entenderlos es sumamente sencillo, debido a que se asemejan a los procesos naturales que poseen las especies para su supervivencia, aplicarlos a un problema real implica la construcción de un sistema de alta complejidad. El presente trabajo contiene el estudio, desarrollo, implementación y prueba de un sistema que aplica Algoritmos Evolucionarios para resolver el problema de la asignación de recursos humanos a tareas vinculadas a la construcción de cañerías para fluidos relacionados con el petróleo.; Genetic algorithms and its generalization: evolutionary algorithms present an alternative to solve complex problems of optimization. If it is easy to understand them, because they are similar to the natural processes that the species possess for its survival, to apply them to a real problem implies the construction of a system of high complexity. This work presents the study, development, implementation, and testing of a system that applies evolutionary algorithms, to solve the problem about assigning human resources to tasks linked to the construction of pipings for fluids related to oil.

2007-01-01T00:00:00Z Los algoritmos Genéticos, y su generalización: los Algoritmos Evolucionarios; presentan una alternativa para solucionar problemas de optimización complejos. Si bien entenderlos es sumamente sencillo, debido a que se asemejan a los procesos naturales que poseen las especies para su supervivencia, aplicarlos a un problema real implica la construcción de un sistema de alta complejidad. El presente trabajo contiene el estudio, desarrollo, implementación y prueba de un sistema que aplica Algoritmos Evolucionarios para resolver el problema de la asignación de recursos humanos a tareas vinculadas a la construcción de cañerías para fluidos relacionados con el petróleo. Genetic algorithms and its generalization: evolutionary algorithms present an alternative to solve complex problems of optimization. If it is easy to understand them, because they are similar to the natural processes that the species possess for its survival, to apply them to a real problem implies the construction of a system of high complexity. This work presents the study, development, implementation, and testing of a system that applies evolutionary algorithms, to solve the problem about assigning human resources to tasks linked to the construction of pipings for fluids related to oil. Rueda, Sonia Vivian Simari, Guillermo Ricardo http://sedici.unlp.edu.ar:80/handle/10915/23374 2018-12-22T20:05:07Z 2007-01-01T00:00:00Z

Objeto de conferencia XIII Congreso Argentino de Ciencias de la Computación Este trabajo propone un modelo para representar el conocimiento y el comportamiento de un agente BDI que actúa e interactúa condicionado por las normas y la estructura social de la organización a la que pertenece. Parte del conocimiento está compartido globalmente y algunos elementos cognitivos son específicos de cada individuo en particular. Cuando el conocimiento de un agente es insuficiente para alcanzar su meta comprometida inicia un diálogo colaborativo. La autonomía y reactividad le permiten adaptarse a las características dinámicas del entorno. La estructura normativa y la asignación de roles brindan patrones para anticipar la evolución de los diálogos.

2007-01-01T00:00:00Z Este trabajo propone un modelo para representar el conocimiento y el comportamiento de un agente BDI que actúa e interactúa condicionado por las normas y la estructura social de la organización a la que pertenece. Parte del conocimiento está compartido globalmente y algunos elementos cognitivos son específicos de cada individuo en particular. Cuando el conocimiento de un agente es insuficiente para alcanzar su meta comprometida inicia un diálogo colaborativo. La autonomía y reactividad le permiten adaptarse a las características dinámicas del entorno. La estructura normativa y la asignación de roles brindan patrones para anticipar la evolución de los diálogos. Odorico, Pablo Delrieux, Claudio http://sedici.unlp.edu.ar:80/handle/10915/23373 2018-12-23T04:02:36Z 2007-01-01T00:00:00Z

Objeto de conferencia XIII Congreso Argentino de Ciencias de la Computación Presentamos un algoritmo para la detección supervisada en tiempo real del horizonte en imágenes de video. El mismo está diseñado para ser implementado en dispositivos autónomos de bajo costo y capacidad de cómputo (microcontroladores), como para ser incorporados en vehículos voladores no tripulados (UAVs). Esto permitiría la obtención de información indispensable para la orientación, navegación, georeferenciamiento y la elaboración de mosaicos de imágenes. El algoritmo se basa en una función de costo para los pixels de un submuestreo de la imagen en las coordenadas YI del espacio cromático YIQ. Con una baja cantidad de evaluaciones el algoritmo determina la orientación más probable del horizonte, y luego refina la búsqueda con un método adaptativo. La implementación de un prototipo en una PC muestra que el algoritmo es lo suficientemente robusto como para determinar la posición del horizonte con muy bajo costo computacional.

2007-01-01T00:00:00Z Presentamos un algoritmo para la detección supervisada en tiempo real del horizonte en imágenes de video. El mismo está diseñado para ser implementado en dispositivos autónomos de bajo costo y capacidad de cómputo (microcontroladores), como para ser incorporados en vehículos voladores no tripulados (UAVs). Esto permitiría la obtención de información indispensable para la orientación, navegación, georeferenciamiento y la elaboración de mosaicos de imágenes. El algoritmo se basa en una función de costo para los pixels de un submuestreo de la imagen en las coordenadas YI del espacio cromático YIQ. Con una baja cantidad de evaluaciones el algoritmo determina la orientación más probable del horizonte, y luego refina la búsqueda con un método adaptativo. La implementación de un prototipo en una PC muestra que el algoritmo es lo suficientemente robusto como para determinar la posición del horizonte con muy bajo costo computacional. Cervino, Claudio Ierache, Jorge Salvador Nasini, Víctor http://sedici.unlp.edu.ar:80/handle/10915/23372 2018-12-23T04:02:30Z 2007-01-01T00:00:00Z

Objeto de conferencia XIII Congreso Argentino de Ciencias de la Computación Se presenta en el contexto de la telemedicina el desarrollo de un sistema de monitoreo a distancia de una unidad electroneumática para la Asistencia Ventricular Izquierda, se exploran los resultados particulares con relación al desarrollo del sistema de monitoreo remoto, y los parámetros fisiológicos del paciente que el sistema monitorea: ECG, presión arterial, temperatura corporal, funcionamiento de la unidad electroneumática, entre estos los valores programados de presión neumática, temperatura de los compresores, nivel de sincronización. En otro orden el trabajo aporta las características de una arquitectura para la integración dinámica del monitoreo de pacientes.; The development of an electropneumatic unit remote monitoring system for the Left Ventricular Assistance appears in the context of telemedicine, the particular results in relation to the development of the remote monitoring system, and the patient´s physiological parameters it monitors: ECG, arterial pressure, corporal temperature, electropneumatic unit operation, and among these ones the programmed values of pneumatic pressure, compressor temperature, and level of synchronization. On the other hand the work involves the characteristics of an architecture for the dynamic integration of the patient´s monitoring.

2007-01-01T00:00:00Z Se presenta en el contexto de la telemedicina el desarrollo de un sistema de monitoreo a distancia de una unidad electroneumática para la Asistencia Ventricular Izquierda, se exploran los resultados particulares con relación al desarrollo del sistema de monitoreo remoto, y los parámetros fisiológicos del paciente que el sistema monitorea: ECG, presión arterial, temperatura corporal, funcionamiento de la unidad electroneumática, entre estos los valores programados de presión neumática, temperatura de los compresores, nivel de sincronización. En otro orden el trabajo aporta las características de una arquitectura para la integración dinámica del monitoreo de pacientes. The development of an electropneumatic unit remote monitoring system for the Left Ventricular Assistance appears in the context of telemedicine, the particular results in relation to the development of the remote monitoring system, and the patient´s physiological parameters it monitors: ECG, arterial pressure, corporal temperature, electropneumatic unit operation, and among these ones the programmed values of pneumatic pressure, compressor temperature, and level of synchronization. On the other hand the work involves the characteristics of an architecture for the dynamic integration of the patient´s monitoring. Villagra, Andrea Pandolfi, Daniel http://sedici.unlp.edu.ar:80/handle/10915/23371 2018-12-23T04:02:25Z 2007-01-01T00:00:00Z

Objeto de conferencia XIII Congreso Argentino de Ciencias de la Computación Los algoritmos de clustering de tipo c-means son sensibles a los valores de inicialización de los centroides y pueden quedar atrapados en extremos locales. Planteado en estos términos, el uso de enfoques aproximados para obtener los centroides más adecuados puede ser de gran utilidad como herramienta complementaria durante ciertas fases del proceso de minería de datos, y en particular dentro de las tareas típicas de minería de datos, entre ellas la de clustering o agrupamiento. En esta dirección, los Algoritmos Genéticos (AGs) y la Optimización Basada en Cúmulo de Partículas (PSO)1 son dos técnicas metaheurísticas poblacionales que podrían utilizarse en este ámbito, más aún cuando los problemas pueden ser planteados como de optimización. En este trabajo se analiza el uso estas dos técnicas metaheurísticas para optimizar la inicialización de los valores de centroides en las funciones aplicadas en los algoritmos de clustering tipo c-means. Los respectivos resultados son comparados usando varios conjuntos de datos generados artificialmente.; The clustering algorithms like c-means are sensitive to the initialization values of the cluster centers and can be trapped by local extrema. In these terms, the use of estimated approaches to obtain the most appropriate cluster centers can be of great utility as a complementary tool during certain phases of the process of data mining; particulary, in some specific task of data mining, e.g., clustering. In this way, Genetic Algorithms (GA) and Particle Swarm Optimization (PSO) are two population metaheuristic approaches that could be considered as optimization. In this work the use of these two metaheuristic approaches is analyzed to optimize the initialization of the cluster centers values in the functions applied in the c-means algorithms. The respective results are compared using several datasets artificially generated.

2007-01-01T00:00:00Z Los algoritmos de clustering de tipo c-means son sensibles a los valores de inicialización de los centroides y pueden quedar atrapados en extremos locales. Planteado en estos términos, el uso de enfoques aproximados para obtener los centroides más adecuados puede ser de gran utilidad como herramienta complementaria durante ciertas fases del proceso de minería de datos, y en particular dentro de las tareas típicas de minería de datos, entre ellas la de clustering o agrupamiento. En esta dirección, los Algoritmos Genéticos (AGs) y la Optimización Basada en Cúmulo de Partículas (PSO)1 son dos técnicas metaheurísticas poblacionales que podrían utilizarse en este ámbito, más aún cuando los problemas pueden ser planteados como de optimización. En este trabajo se analiza el uso estas dos técnicas metaheurísticas para optimizar la inicialización de los valores de centroides en las funciones aplicadas en los algoritmos de clustering tipo c-means. Los respectivos resultados son comparados usando varios conjuntos de datos generados artificialmente. The clustering algorithms like c-means are sensitive to the initialization values of the cluster centers and can be trapped by local extrema. In these terms, the use of estimated approaches to obtain the most appropriate cluster centers can be of great utility as a complementary tool during certain phases of the process of data mining; particulary, in some specific task of data mining, e.g., clustering. In this way, Genetic Algorithms (GA) and Particle Swarm Optimization (PSO) are two population metaheuristic approaches that could be considered as optimization. In this work the use of these two metaheuristic approaches is analyzed to optimize the initialization of the cluster centers values in the functions applied in the c-means algorithms. The respective results are compared using several datasets artificially generated. Villagra, Andrea Montenegro, Cristian San Pedro, María Eugenia de Lasso, Marta Graciela Pandolfi, Daniel http://sedici.unlp.edu.ar:80/handle/10915/23370 2018-12-23T04:02:13Z 2007-01-01T00:00:00Z

Objeto de conferencia XIII Congreso Argentino de Ciencias de la Computación En este trabajo se presenta una herramienta (PAE) que utiliza un algoritmo evolutivo, generador de múltiples soluciones, para la replanificación dinámica del mantenimiento preventivo de locaciones petroleras. La explotación y el transporte de petróleo son actividades muy importantes para el desarrollo económico de la sociedad industrial moderna. Sin embargo, estas actividades son generadoras de riesgos que se traducen en contaminaciones. Es importante que las empresas petroleras realicen un correcto mantenimiento de sus locaciones. PAE brinda en forma oportuna la planificación del recorrido y su benefic io debe observarse desde dos aspectos. Primero, una planificación es mejor que otra, si para un mismo número de locaciones a visitar el costo de recorrido e intervención planificado es menor. Segundo, si con un mismo tiempo de intervención es posible realizar el mantenimiento a más locaciones, este beneficio reduce la probabilidad de caída de la producción al incrementar la cantidad de locaciones recorridas. PAE está en la cuarta fase de su desarrollo, donde se han incorporado restricciones al momento de la replanificación. Los resultados obtenidos han sido satisfactorios ya que minimizan el tiempo total de una planificación y maximizan la cantidad de locaciones visitadas, satisfaciendo las restricciones.; This work deals with an application (PAE) that uses an evolutionary algorithm, generator of multiple solutions for the rescheduling of preventive maintenance of oil fields. The exploitation and the transport of oil are very important activities for the economic development of the industrial modern society. However, these activities are generating risks that are translated in contaminations. It is important that oil companies carry out a correct maintenance of their oil fields. PAE provides a maintenance visit schedule at the right moment and this benefit should be observed from two aspects. First, a schedule is better than other if for the same number of oil wells to be visited, the round cost and the time scheduled is small. Second, if for the same period of time it is possible to carry out the maintenance to more oil wells, this benefit reduces the probability of falling when the number of visited oil wells is increased. PAE is in the fourth phase of its development, in which we have incorporated restrictions to the moment of rescheduling. The obtained results have been satisfactory since they minimize the total time of a scheduling and they maximize the quantity of visited oil wells, satisfying the constraints.

2007-01-01T00:00:00Z En este trabajo se presenta una herramienta (PAE) que utiliza un algoritmo evolutivo, generador de múltiples soluciones, para la replanificación dinámica del mantenimiento preventivo de locaciones petroleras. La explotación y el transporte de petróleo son actividades muy importantes para el desarrollo económico de la sociedad industrial moderna. Sin embargo, estas actividades son generadoras de riesgos que se traducen en contaminaciones. Es importante que las empresas petroleras realicen un correcto mantenimiento de sus locaciones. PAE brinda en forma oportuna la planificación del recorrido y su benefic io debe observarse desde dos aspectos. Primero, una planificación es mejor que otra, si para un mismo número de locaciones a visitar el costo de recorrido e intervención planificado es menor. Segundo, si con un mismo tiempo de intervención es posible realizar el mantenimiento a más locaciones, este beneficio reduce la probabilidad de caída de la producción al incrementar la cantidad de locaciones recorridas. PAE está en la cuarta fase de su desarrollo, donde se han incorporado restricciones al momento de la replanificación. Los resultados obtenidos han sido satisfactorios ya que minimizan el tiempo total de una planificación y maximizan la cantidad de locaciones visitadas, satisfaciendo las restricciones. This work deals with an application (PAE) that uses an evolutionary algorithm, generator of multiple solutions for the rescheduling of preventive maintenance of oil fields. The exploitation and the transport of oil are very important activities for the economic development of the industrial modern society. However, these activities are generating risks that are translated in contaminations. It is important that oil companies carry out a correct maintenance of their oil fields. PAE provides a maintenance visit schedule at the right moment and this benefit should be observed from two aspects. First, a schedule is better than other if for the same number of oil wells to be visited, the round cost and the time scheduled is small. Second, if for the same period of time it is possible to carry out the maintenance to more oil wells, this benefit reduces the probability of falling when the number of visited oil wells is increased. PAE is in the fourth phase of its development, in which we have incorporated restrictions to the moment of rescheduling. The obtained results have been satisfactory since they minimize the total time of a scheduling and they maximize the quantity of visited oil wells, satisfying the constraints.