Abstract

Содержание

Introduction

Modern society is acquiring more and more features of information. At the same time, the information society requires a new, better level of education and new teaching methods. These requirements are due to the large involvement of people in processes that require high narrow professional education, and there is a constant need for retraining of workers, as technologies are developing rapidly [1].

Today, higher education institutions are no longer able to quickly change courses and respond quickly to changes in the demands of consumers of educational services. Modern production requires more educated people than 10 years ago. The experience of universities in Western countries shows that the University is no longer tied to the area (innovative universities have abandoned geographical names). In addition, the concept of cyclical studies disappears: you can join some schools on any day. Instead of the faculty, there is now a list of courses required for qualification, the University becomes a global organization, and in the center of the direct educational process is no longer a Professor who gathers an audience around him, but a student who is served by professors [1].

1. Theme urgency

Open education in many countries is considered today as a system that provides nationwide access to educational resources through the widespread use of information and educational technologies of distance education. And on this basis, it provides conditions for the most complete realization of citizens ' rights to education, which in structure and quality meets the requirements of economic development and civil society [1].

Against the background of these trends, society's demands for education have changed markedly. Distance education systems such as" Virtual University "or" Virtual Department " are becoming relevant. Modern information technologies allow satisfying the demands of the society [1].

This master's work is just devoted to such an urgent scientific task of developing a virtual Department of student learning, based on the agent system.

2. Goal and tasks of the research

The aim of the study is to design and develop the Department of education of the University as an object of virtual educational environment [2], which is an information space of interaction between the participants of the educational process generated by information and communication technologies, including a set of computer tools and technologies that allows to manage the content of the educational environment and communication of participants.

Master's Problems:

  1. Perform agent-oriented analysis of the learning process of students at the Cathedral level.
  2. Get a new model of individual training of students in a particular discipline.
  3. To Develop the structure of the agent-oriented model of the Department.
  4. Implement a multi-agent system using the Jack tool environment.

Object of study: modeling of a multi-agent system.

Subject of study: multi-agent system and intelligent agents.

Scientific novelty is to create a model of individual learning, to preserve the relationship between the participants of the educational process, close to the real-life, and provided an opportunity for Autonomous and distributed performance of educational and methodical duties. This model will improve the decentralization and individuality of teachers and students at the Department level [3].

Practical results it is planned to obtain a ready virtual Department, presented in the form of a multi-agent software system based on the principles of distributed artificial intelligence. The advantage of the virtual Department [4] as a new learning environment is as follows:

  1. Flexibility-the student has the opportunity to study in a convenient place, because the learning cycle is carried out through Internet technologies;
  2. Modularity-the student has the opportunity to form a curriculum that meets individual needs from a set of independent courses-modules;
  3. Economic efficiency – the costs of both the student and the education system are reduced through the most efficient use of educational space, time and technical means [4].

3. Research and development review

Multi-Agent systems nowadays are widely used in various fields such as mobile networks, web, planning and logistics, distributed systems, research projects, industry and engineering systems modeling [5].

Since the mid-1990s, agent-based models have been used to solve many commercial and technological problems [6]. Examples are the following tasks:

  1. The optimization of supply chain and logistics.
  2. Modeling of consumer behavior (including social networks).
  3. Distributed computing.
  4. Human resources Management.
  5. Transport Management.
  6. Portfolio Management [6].

3.1 Overview of international sources

the First agent-based model was developed in the late 1940s. Subsequently, the development of microcomputers has led to the further development of this direction and the capacity to conduct computer simulations. It is believed that agent-oriented models originate from the computers of John von Neumann (Von Neumann), which are theoretical machines capable of reproduction.

the use of an agent-based model for social systems began with the work of the programmer Craig Reynolds, in which he attempted to model the activities of living biological agents (the "Artificial life" model) [6].

there Are various organizations involved in the study of multi-agent systems [7 - 9].

3.2 Overview of national sources

Russian scientists are also actively engaged in research and development of multi-agent systems. The most outstanding representatives are Tarasov V. B. [10-12], who investigated the features of neural networks and intelligent agents, Karsaev O. V., Groom G. V., Samoilov V. V. studying the development environment and language for development of applied multi-agent systems [13]. Also, the study of agent-oriented agents was carried out by Krizhanovsky, A. A. [14] and Skobelev, p. O. [15].

3.3 Overview of local sources

at Donetsk national technical University (Department of computer engineering) my supervisor Oleg Fedyaev is engaged in the study of multi-agent systems. He began to study agent-oriented systems and work on the creation of a virtual Department of the University together with Zhabskaya I.e. [16].


In the future continued engaged in the research of multiagent systems along with a master's of previous years:

  1. Lyamin R.V. Multi-Agent system of teaching students at the Cathedral level.
  2. Zaitsev I.M. Models of collective behavior of intelligent agents in multi-agent systems of modeling and enterprise management.
  3. Zudikova Y.V. Assessment of the effectiveness of multi-agent simulation systems with distributed intelligence.
  4. Lukina Y.Y. Agent-oriented software models of human behavior in the socio-economic environment.
  5. Strogalov A.S. Neural network model of software agents in social–oriented multi-agent systems.
  6. Grabchuk O.P. Agent-oriented modeling of training and employment of young specialists.
  7. Eliferov V.V. Multi-Agent simulation model for predicting the results of training and employment of specialists.
  8. Kutashov R.I. Software implementation of agent-oriented system of distance learning of students in technical disciplines.
  9. Medhaus S.V. Architecture and functioning of software agents in simulation models of employment of graduates of the University a multi-agent type.

4. Construction of logical models of virtual Department

To develop a multi-agent application that automates the learning process at the departmental level, an agent-oriented analysis of the subject area of the educational process was carried out according to the Gaia methodology [17]. Using this methodology, the models necessary to describe the virtual Department and the subsequent software implementation [18] are developed.

Learning Process, according to Gaia methodology, is described by the following models: role model, interaction model, agent model, service model, communication model. The relationships and content of the models obtained with the help of agent-oriented design of the virtual Department are presented below (see Fig. 1 [18]).

Relationship of models in agent-oriented design of the learning process

Figure 1 - Relationship of models in agent-oriented design of the learning process

At the stage of agent-based analysis and design, the following models have been developed: a role model to describe the job responsibilities of all roles in the form of activities, protocols, authorities, obligations; a model of interactions to describe the main types of communication between roles in the form of protocols; a model of agents to determine the types of agents; a model of functioning to determine the actions of agents and a model of communications to display possible communications between agents.

Since a role is an abstract description of the functions of an official, each role is characterized by independent actions (activities), interactions with other roles (communication), authority over the resources required to perform it, and obligations that define the "life cycle" of the role and are described by regular expressions. The purpose of the role model is to formally describe the content of all roles in the organization [18].

The agent Model is used to define the agent types used in the system. Similar roles are combined into one agent type. The decision to combine multiple roles into one agent type improves the understanding of the functional purpose of agents and contributes to the effectiveness of their software implementation.

The operating model defines for each agent the actions to be performed by them in accordance with the viability obligations of the respective role. For each function performed by the agent, the functioning model defines the input and output data, pre-and post-conditions, depending on which the agent initiates execution and determines the completion of the function [18].

The communication model reflects possible communication links between agents. It is based on the role model, interaction model, and agent model. The model of relations is presented in the form of a directed graph. Nodes correspond to agent types and arcs correspond to communication links. Communications between agents involve the transmission of messages between agents in both directions. On its basis, for each agent, a list of agents with which it is possible to establish unilateral or bilateral relations [18].

Thanks to the developed agent-oriented models, a systematic transition from the stage of setting the problem to the stage of software implementation of the computer environment with elements of communication between the subjects of the educational process of studying the disciplines of the Department.

5. Connection of logical models with visual models in cross-platform environment Jack

JACK IS a professional, cross-platform environment for the creation, operation and integration of commercial agro-industrial systems. It is built on a solid logical Foundation: a model of beliefs, desires and intentions (belief, desire, and intention (BDI) model). BDI is an intuitive and powerful abstraction that allows developers to manage the complexity of a problem. In JACK, agents are defined in terms of their beliefs (what they know and what they know how to do), their desires (what goals they want to achieve) and their intentions (the goals they currently pursue to achieve).

The BDI architecture allows you to model the mental properties of the agents needed to solve the learning problem. An agent having a BDI architecture is described by three components a = (B, D, I), where B are the agent's beliefs, which are the agent's information about his own state and the state of his environment, and are considered as his information component; D are the agent's desires in the form of information about his goals, which are considered as his motivational component; I is the intention of the agent, which represents the possible course of action, and is its judicious component [19].

To programmatically implement the agent's beliefs, desires, and intentions, JACK provides the following new constructs that extend the Java language syntax at the class level:

  1. Agent, defines intelligent agents.
  2. Event, defines agent targets, in the form of events, to simulate situations and messages that the agent should be able to respond to.
  3. Plan, describes the intentions of the agent to achieve the goal in the form of plans and conditions of their applicability.
  4. Beliefset, describes the knowledge of the agent.
  5. Capability, structures beliefs, events, and plans into clusters to implement a specific agent's intellectual ability to achieve a goal [20].

The semantics of the visual model of the agent presented formally in the JACK environment in the form of Аgent = (N, Bel, PE, HE, SE, PS), is described as follows: N – the name of the agent; Bel – beliefs of the agent; PE = {E1, E2, ..., En} is the set of event names, generated its own methods agent; HE = {{E1, E2, ..., En}, HE1, ...} is the set of events to be processed, independently generated and perceived from the outside; SE = {SE1, SE2, ..., SEm} is the set of event names sent by the agent to the external medium; PS = {P1, P2, ..., Pk} is the set of plan names that define the behavior of the agent [20].

Highlighting the most important aspects of the semantics of visual models of environment JACK ensure the quality of their build.

In order to work in the Jack tool environment, you need to think at the level of ITS concepts specific to the BDI architecture. For the correct transition from abstract models of agents to their representation at the level of visual models of JACK environment, the authors have compiled specifications of semantics of visual models of this environment [20].

The physical components of the agents are represented as classes with specific relationships between them (see Fig. 2 [20]).

Student software agent Architecture in JACK system

Figure 2 - Student software agent Architecture in JACK system

Conclusions

Master's thesis is devoted to the actual scientific problem of designing and developing the educational Department of the University as an object of virtual educational environment. Within the framework of the conducted studies the following was performed:

  1. Agent-oriented analysis of the process of teaching students at the Cathedral level, on the basis of which a new model of the Department of the University.
  2. A new model of individual training of students in a particular discipline has been Obtained.
  3. The structure of the agent-oriented model of the Department was Developed.
  4. In an instrumental environment, which are modeled by agents, is used software environment Jack.

When writing this essay master's work has not yet been completed. Final completion: may 2019. The full text of the work and materials on the topic can be obtained from the author or his supervisor after the specified date.

References

  1. Виртуальный Университет – О системе дистанционного образования «Виртуальный Университет» [Электронный ресурс]. – Режим доступа: http://vu.net.ua/ru/.
  2. Studexpo – Виртуальная образовательная среда [Электронный ресурс]. – Режим доступа: https://studexpo.ru/214575/pedagogika/virtualnaya_obrazovatelnaya_sreda.
  3. Сб. науч. тр. ДонНТУ. Серия: «Проблемы моделирования и автоматизации проектирования динамических систем»(МАП-2006), выпуск 5(116) — Донецк: ДонНТУ, 2006. — с. 105-116. http://ea.donntu.ru/bitstream/123456789/5815/1/11.pdf.
  4. Бобродобро – Преимущества виртуальной образовательной среды [Электронный ресурс]. – Режим доступа: http://ped.bobrodobro.ru/3170.
  5. Мультиагентные системы — что это [Электронный ресурс]. – Режим доступа: https://habr.com/post/70446/.
  6. Wiki – Агентное моделирование [Электронный ресурс]. – Режим доступа: https://ru.wikipedia.org/wiki/%D0%90%D0%B3%D0%B5%D0%BD%D1%82%D0%BD%D0%BE%D0%B5_%D0%BC%D0%BE%D0%B4%D0%B5%D0%BB%D0%B8%D1%80%D0%BE%D0%B2%D0%B0%D0%BD%D0%B8%D0%B5.
  7. Multiagent Systems Research Group - Critical MAS [Электронный ресурс]. – Режим доступа: http://www.cs.huji.ac.il/labs/mas/proc.php.
  8. Multi-Agent Systems Lab - Лаборатория многоагентных систем [Электронный ресурс]. – Режим доступа: http://mas.cs.umass.edu/.
  9. Fipa - организация стандартизации агентных технологий [Электронный ресурс]. – Режим доступа: http://www.fipa.org/index.html.
  10. Тарасов В.Б. Системно-организационный подход в искусственном интеллекте / В.Б. Тарасов // Программные продукты и системы. – 1999 - №3. – С.6-13.http://is.ifmo.ru/art_int/_sopii.pdf.
  11. Тарасов В.Б. Агенты, многоагентные системы, виртуальные сообщества: стратегическое направление в информатике и искусственном интеллекте / В.Б. Тарасов // Новости искусственного интеллекта. – 1998. – №2. – С.1-46. http://science.donntu.ru/ius/kirgaev/library/Tarasov-vb-agenty-mnogoagentnye-sistemy-virtualnye-soobschestva-strategicheskoe-napravlenie-v-informatike-i-iskusstvennom-intellekte.pdf.
  12. Тарасов В.Б. Моделирование мнений агентов в многоагентных системах на основе четырехзначных семантик / В.Б. Тарасов, А.В. Борисов // Программные продукты и системы. – 2006. – №2. – С.47-50. https://cyberleninka.ru/article/v/modelirovanie-mneniy-agentov-v-mnogoagentnyh-sistemah-na-osnove-chetyrehznachnyh-semantik.
  13. Карсаев О.В. Инструментальная среда и язык для разработки прикладных многоагентных систем / О.В. Карсаев, В.Г. Конюший, В.В. Самойлов // Известия ЮФУ. Технические науки. – 3 мая 2010. – С.24-28. http://izv-tn.tti.sfedu.ru/?p=14216.
  14. Крижановский А.А. Вопросы реализации проблемно-ориентированных агентов интеграции знаний / А.А. Крижановский // Труды СПИИРАН, Том 1. - СПб: СПИИРАН, 2001.
  15. Скобелев П.О. Открытые мультиагентные системы для поддержки процессов принятия решений при управлении предприятиями / П.О. Скобелев // Известия Самарского научного центра Российской академии наук – № 1 – Том 3. – 2001.
  16. Жабская Т.Е. Разработка агентно-ориентированной интеллектуальной среды обучения / Т.Е. Жабская, О.И. Федяев // Материалы четвертой межнародной научно-технической конференции «Моделирование и компьютерная графика» — 5-8 октября 2011 г., Донецк, ДонНТУ, 2011. — с. 102–108. http://ea.donntu.ru:8080/bitstream/123456789/3005/1/c_17.pdf.
  17. F.Zambonelli Developing Multiagent Systems: The Gaia Methodology / Zambonelli F., Jennings N.R., Wooldridge M. // In ACM Transactions on Software Engineering Methodology, 12(3): июль 2003. — с. 317-370.
  18. Жабская Т.Е. Архитектура и функционирование программных агентов в обучающей системе многоагентного типа / Т.Е. Жабская, О.И. Федяев // Научн. тр. ДонНТУ, серия «Информатика, кибернетика и вычислительная техника» (ИКВТ-2008), выпуск 9(132) — Донецк: ДонНТУ, 2008. — с. 249-254. http://irbis-nbuv.gov.ua/cgi-bin/irbis_nbuv/cgiirbis_64.exe?I21DBN=LINK&P21DBN=UJRN&Z21ID=&S21REF=10&S21CNR=20&S21STN=1&S21FMT=ASP_meta&C21COM=S&2_S21P03=FILA=&2_S21STR=Npdntu_inf_2008_9_42.
  19. Kinny D. A Methodology and Modelling Technique for Systems of BDI Agents / D. Kinny, M. Georgeff, A. Rao // Proceedings of the Seventh European Workshop on Modelling Autonomous Agents in a Multi-Agent MAAMAW, 1996, (LNAI Volume 1038).
  20. Жабская Т.Е. Проектирование виртуальной кафедры университета на основе многомодельного агентно-ориентированного подхода / Т.Е. Жабская, О.И. Федяев // Искусственный интеллект — 2010, №3. — с. 679–686. http://dspace.nbuv.gov.ua/bitstream/handle/123456789/56853/83-Fedyaev.pdf?sequence=1.