Patrones de implementación para incluir comportamientos proactivos

Espino, Mailyn Moreno; Carrasco, Alternán; Rosete, Alejandro; Delgado, Martha D.

doi:10.17562/pb-47-9

Cited by 2 publications

(2 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Los agentes que cumplirán con estos roles tendrán la capacidad de comunicarse y actuar autónomamente, pudiendo hasta cambiar su comportamiento en dependencia de las situaciones a las que se enfrenten [7].…”

Section: Patrón Proactive Observer_jadeunclassified

See 1 more Smart Citation

Transformación del Q-Learning para el Aprendizaje en Agentes JADE

Cepero-Pérez¹,

Espino

2015

Lámpsakos

View full text Add to dashboard Cite

Resumen. El aumento de la interacción entre los sistemas informáticos ha modifi cado la forma tradicional de analizarlos y desarrollarlos. La necesidad de la interacción entre los componentes del sistema es cada vez más importante para poder resolver tareas conjuntas, que de forma individual serían muy costosas o incluso imposibles de desarrollar. Los sistemas multi-agente ofrecen una arquitectura interesante y completa para ejecutar tareas distribuidas que cooperan entre sí. La creación de un sistema multiagente o un agente requiere de gran esfuerzo, por lo que se han adoptado métodos como los patrones de implementación. El patrón Proactive Obsever_ JADE permite crear los agentes e incluirle en cada uno comportamientos dotados de inteligencia que pueden evolucionar utilizando técnicas de aprendizaje automático. El aprendizaje por refuerzo es una técnica del aprendizaje automático que permite a los agentes aprender a través de interacciones de prueba y error, en un ambiente dinámico. El aprendizaje por refuerzo en sistemas multi-agente ofrece nuevos retos derivados de la distribución del aprendizaje, como pueden ser la necesidad de la coordinación entre agentes o la distribución del conocimiento, que deben ser analizados y tratados.Palabras clave: agentes, aprendizaj e reforzado, JADE, patrones de implementación. Abstract. Increased interaction between computersystems has modifi ed the traditional way to analyze and develop them. The need for interaction between the system components is increasingly important to solve joint tasks, which individually would be very expensive or even impossible to develop once. Multiagent systems offer an interesting and complete distributed architecture to execute tasks cooperate. The creation of a multi-agent system or an agent requires great effort so methods have been adopted as the deployment patterns. The pattern creates Proactive Obsever_JADE agents and include in each endowed with intelligence behaviors can evolve using machine learning techniques. The reinforcement learning is a machine learning technique that allows agents to learn through trial and error interactions in a dynamic environment. Reinforcement learning in multi-agent systems offers new challenges arising from the distribution of learning, such as the need for coordination between agents or distribution of knowledge, which should be analyzed and treated.

show abstract

Section: Patrón Proactive Observer_jadeunclassified

“…Con el patrón Proactive Observer_JADE se facilita el trabajo con la plataforma de desarrollo JADE [7], se incorporan comportamientos proactivos; además se permite al usuario realizar cambios en el comportamiento de los agentes de forma dinámica, lo que es un aspecto primordial en el desarrollo de un SMA [8].…”

Section: Introductionunclassified

Transformación del Q-Learning para el Aprendizaje en Agentes JADE

Cepero-Pérez¹,

Espino

2015

Lámpsakos

View full text Add to dashboard Cite

show abstract

Goal-oriented Email Stream Classifier with A Multi-agent System Approach

Hojas-Mazo¹,

Espino²,

Martínez³

et al. 2021

IJACSA

View full text Add to dashboard Cite

Now-a-days, email is often one of the most widely used means of communication despite the rise of other communication methods such as instant messaging or communication via social networks. The need to automate the email stream management increases for reasons such as multifolder categorization, and spam email classification. There are solutions based on email content, capable of contemplating elements such as the text subjective nature, adverse effects of concept drift, among others. This paper presents an email stream classifier with a goal-oriented approach to client and server environment. The i* language was the basis for designing the proposed email stream classifier. The email environment was represented with the early requirements model and the proposed classifier with the late requirements model. The classifier was implemented following a multi-agent system approach supported by JADE agent platform and Implementation_JADE pattern. The behavior of agents was taking from an existing classifier.The multi-agent classifier was evaluated using functional, efficacy and performance tests, which compared the existing classifier with the multi-agent approach. The results obtained were satisfactory in all the tests. The performance of multi-agent approach was better than the existing classifier due to the use of multi-threads.

show abstract

Patrones de implementación para incluir comportamientos proactivos

Cited by 2 publications

References 20 publications

Transformación del Q-Learning para el Aprendizaje en Agentes JADE

Transformación del Q-Learning para el Aprendizaje en Agentes JADE

Goal-oriented Email Stream Classifier with A Multi-agent System Approach

Contact Info

Product

Resources

About