Model of a multi­agent schedule planning system for air traffic controllers

DOI: 10.31673/2412-9070.2020.020812

Abstract

The article proves that the approach to the development of planning and scheduling systems using multi-agent technology is currently considered by specialists as one of the most promising. This article developed a model of such a system with an emphasis on the development of a distributed knowledge base, the components of which are the knowledge bases of intelligent agents. Knowledge agents are divided into formal and heuristic. Formal knowledge corresponds to the knowledge that sets the constraints of the problem. Heuristic knowledge partially reflects the subjective experience of experts, and partially is the result of the analysis and generalization of their positive and negative experience obtained on the basis of training the system on precedents. The formalization of heuristic knowledge has additional restrictions on possible protocols for the interaction of agents, which are based on a model of controlled «auction». The main problem in the development of distributed planning systems is the creation of a distributed knowledge base and interaction models of agents supporting the distributed use of knowledge in the process of the system.
The article discusses the developed technology for creating multi-agent planning and scheduling systems, based on the developed tool to support the processes of creating multi-agent systems. A significant number of practically important applied tasks in the automation of the educational process is reduced to combinatorial formulation. First of all, these include tasks whose formal formulation is reduced to a planning and scheduling model in conditions of limited resources and real-time constraints. This model corresponds to a wide range of fairly traditional tasks of scheduling. Given the changing approaches to training in connection with the introduction of quarantine, the task of compiling effective class schedules is very important. Recently, in connection with new trends in the use of network and information technologies in the field of organization of the educational process, performances have appeared that can take into account the level of knowledge, workload of the simulator and training material base, taking into account the individual learning path, training in groups for planning schedules.

Keywords: air traffic control dispatcher; intelligent training system; information and analytical system; training model; assessment and testing process.

References
1. Маслобоев А. В. Гибридная архитектура интеллектуального агента с имитационным аппаратом // Вестник МГТУ. 2009. №1. С. 67–78.
2. Городецкий В. И., Карасев О. В. Технология разработки прикладных многоагентных систем в инструментальной среде MASDK // Труды СПИИРАН. Вып. 3. Т. 1. СПб.: Наука, 2006. С. 23–54.
3. Bugaychenko D. Y. MASL: A logic for the specification of multiagent real-time systems: Proc. 5th International Central and Eastern European Conference on Multi-Agent Systems. Leipzig (Germany): Springer-Verlag, 2017. P. 183–192.
4. Wooldridge M. Jennings N., and Kinny D. The Gaia Methodology for Agent-Oriented Analysis and Design // International Jour. of Autonomous Agents and Multi-Agent Systems. 2000. №3(3). P. 285–312.
5. Бережний А. О., Сорока М. Ю., Сало Н. А. Методи рішення завдань планування поведінки агентів в інтелектуальних системах підтримки прийняття рішень: зб. наук. праць Харків. нац. ун-ту Повітряних Сил. 2019. № 4(62). С. 18–24.
6. Trystan A. V. Soroka M. Yu. Automation of the educational process in Ukraine higher military education institutions Modern Problems Of Computer Science And IT-Education: collective monograph [editorial board K. Melnyk, O. Shmatko]. Vienna: Premier Publishing s.r.o., 2020. Р. 224–236.
7. Сорока М. Ю., Сало Н. А. Кібербезпека та інформаційні технології: монографія. Харків: ТОВ «ДІСА ПЛЮС», 2020. 380 с.
8. Proceeding of the Fifth International Conference «The Practical Application of Intelligent Agents and Multi-agent Technology» (PAAM’2000), London, UK, 2000.
9. Safra S., Tennenholtz M. On Planning while Learning // Journ. of Artificial Intelligence Research. 1994. No. 9(2). P. 111–129.
10. Левин В. И. Теория расписаний и непрерывная логика. Москва: LAP, 2011. 124 с.
11. Красный Д. Г., Нейдорф Р. А., Кобак В. Г. Исследование неоднородных распределительных задач теории расписаний. Москва: LAP, 2011. 184 с.
12. Сидоркина И. Г. Системы искусственного интеллекта. Москва: КноРус, 2011. 248 с.