Abstract «Models of collective behavior of intelligent agents in modeling of the multi-agent system and enterprise management»

Zaytsev Ivan Mihaylovich

Zaytsev Ivan Mihaylovich

Faculty: Computers and Information Science

Department: Applied Mathematics and Information Science

Speciality: Software engineering

Master's theme: Models of collective behavior of intelligent agents in modeling of the multi-agent system and enterprise management

Scientific advisor: Fedyaev Oleg Ivanovich

Other contact information can be found on my personal site.

Introduction

The «Enterprise» definition includes a wide range of subject areas to demonstrate the capabilities of multi-agent approach, so I had selected the subject area directly related to the educational process to demonstrate the possibilities of the multi-agent environment JADE. The modern university is an established hierarchical structure, which is closely related to the needs of the personal attendance, such as students and teachers. Such restrictions have emerged long before the existence of a global network in its modern form. Clearly, in order to meet the most modern requirements to education, the method must change dynamically. My work is dedicated to one of the possible improvements of the modern university —to the creation of a virtual model of the cathedra with which it would be possible to transfer part of the teachers’ work. The goals and tasks that must be reached The purpose of the research is to study multi-agent systems, the application of agent-oriented approach to the solution of the problem — the creation of a multi-agent virtual cathedra, as well as the implementation of this software using JADE. The relevance of the topic The development of technology of artificial agents, the creation of multi-agent systems (MAS) and the virtual organizations is one of the most important and promising areas of development of the new information and communication technologies.

Nowadays, agent-oriented approach is widely used in such areas as the distributed solutions of complex problems, product design, business reengineering and the construction of the virtual enterprise, simulation of integrated production systems and electronic commerce, organization of robots collective work. The current state of science and education in Ukraine, due to economic factors, is characterized by increase of training quality and determines the constant search for new methods and means of improving the educational process. Distance education — a new form of organization of the educational process, based on the principles of student self-studying. The learning environment is characterized by the fact that students are mostly far from the teacher, but at the same time, they have the opportunity to maintain a dialogue with the means of telecommunication. The development of a virtual cathedra is decided to be an important task of the modern educational process, because it allows them to transfer part of the work of teaching staff to the system, more gives flexibility to interact with students, to use an individual approach.

Alleged scientific novelty

Usage of the intelligent agents provides an individual approach to each student, based not only on the general requirements, but also on the characteristics and knowledge of the student. In addition, the usage of the multi-agent will improve reliability by providing an educational system in the form of a system with distributed artificial intelligence. In the developed multi-agent system I use Jade, that makes the difference from other similar systems using agent approach.

Expected results

After the end of the software system development it is scheduled to get the working software project ready to be implemented in the educational process of the cathedra. It is also planned to conduct a comprehensive analysis and comparison of the practical aspects of implementing multi-agent systems using different software development tools.

The review of the researches and the developments of the theme:

Throughout the world there are developments on such software systems. I would like to list the most successful of them:

  1. KOD (Knowledge On Demand) - a project designed to improve the learning process, using an individual learner approach.
  2. IDEAL. The system supports personalized customer interaction with the educational system, helps to automatically evaluate the results, and to provide easy-to-use tool for develop of the interactive multimedia applications.

At the national level, the system of distance education in Kherson State Technical University. The system of distance education (SDO) which operates in the global computer network Internet, is an environment of so-called "virtual education". [3]

Representatives of the local level:

  1. Lyamin Roman V., "Multi-agent system of students education on the cathedral level" [4]. The main difference between the Roman and my work is the choice of tools. His work uses multi-agent system Bee Gent, while my one is based on JADE.
  2. Hrach Eugene G., «The company Model as an artificial intelligent system for analysis and management on the basis of knowledges» [5]. In this paper, Eugenia is also used multi-agent approach for decision-making and enterprise management.

The theory of the agent-oriented systems

Every agent in modern distributed systems is a process that has enough knowledge about the object and the opportunity to share this knowledge with other agents. In the terms of the object-oriented approach, the agent can be decided as a set of functions in conjunction with an interface that is able to send and receive questions and answers.

At the moment we have a large variety of special software development tools for agent systems [1]. JADE (Java Agent Development Environment) is one of them.

Usage of agents makes it possible to simplify the transitions of the real world, real learning process, to its display in the virtual world. Lokking at the same model of education as a real university, we take the first step to preserve the quality of distance learning.

Agent platform is an middleware executive level, which is located between the agents and the operating system.

JADE agent platform is completely written in Java programming language using such advanced capabilities as Java RMI, Java CORBA IDL, Java Serialization and Java Reflection API.

It simplifies the development of multi-agent systems through the use of FIPA-specifications and also a series of instruments (tools), which supports the phase error correction (debugging) and Deployment (deployment) system.

The number of companies and academic groups are using JADE. Among them there are such well-known companies, as BT, CNET, NHK, Imperial College, IRST, KPN, University of Helsinky, INRIA, ATOS, and many others [2].

Conclusion

On this stage complete analysis of the chosen subject area forms the technical requirements. Database structure, which has been designed satisfyes all the needs. The selected software environment — JADE — provides a platform for agents and is the best suitable system for this kind of development. The development of software system, based on defined requirements, has been started.

References

  1. Городецкий В.И. Многоагентные системы: современное состояние исследований и перспективы развития // Сб. докл. V национальной конференции «Искусственный интеллект — 96», т.1 — Казань, 1996. — с. 36-45.
  2. Глибовец Н.Н. Использование JADE (Java Agent Development Environment) для разработки компьютерных систем поддержки дистанционного обучения агентного типа. // Educational Technology & Society — №8(3) — 2005. — с. 325-345.
  3. Лямин Р. В., Федяев О.И. Многоагентная система обучения студентов на кафедральном уровне. // Сб. науч. тр. ДонНТУ. Серия: «Проблемы моделирования и автоматизации проектирования динамических систем» (МАП-2006), выпуск 5(116) — Донецк: ДонНТУ, 2006. — с. 46-49. [Электронный ресурс] / Masters' Portal ДонНТУ, — http://masters.donntu.ru/2008/fvti/lyamin/diss/index.htm
  4. Грач Е.Г. Модель предприятия как интеллектуальная искусственная система для анализа и управления на основе знаний //Сб. трудов Международной студенческой научно-практической конференции «Информатика и компьютерные технологии». — Донецк: ДонНТУ, 2005. — с. 231-232. [Электронный ресурс] / Masters' Portal ДонНТУ, — http://masters.donntu.ru/2006/fvti/grach/diss/index.htm
  5. Тарасов В.Б. Агенты, многоагентные системы, виртуальные сообщества: стратегическое направление в информатике и искусственном интеллекте // Жур. «Новости искусственного интеллекта» Киев, 1998. — с.21-37