A MODEL OF INTELLIGENT RESOURCE MANAGEMENT OF CLASS MANET TERRESTRIAL COMMUNICATION NETWORK
DOI:
https://doi.org/10.32689/maup.it.2023.3.1Keywords:
communication network, MANET, intelligent control system, neural networks, machine learning with reinforcement, Markov process, predictionAbstract
The article is devoted to the development of a model of intelligent resource management of the ground MANET class communication network. Management of MANET networks is a difficult task due to their dynamic nature, high mobility of nodes, limited resources: battery power, technical characteristics of communication devices, protocols of different levels of the OSI model, and the need to implement management functions at the node and network level in the absence of centralized control. The scientific novelty of the developed mathematical model consists in the introduction of machine learning algorithms with reinforcement for managing the process of forming control decisions at the network level, and a set of neural networks to ensure the requirements for the quality of information exchange at the node level (implementation of user goals). The learning process of neural networks includes the use of a new mobility model that takes into account the physical parameters of network nodes, and in combination with radio connectivity metrics and routing metrics according to the selected protocol, adaptation to changes in the network is ensured in real time. In addition, the use of robotic platforms - mobile base stations can increase the flexibility and adaptability of the network. The use of a two-level intelligent control system allows dividing the process into two stages. At the first stage, it is proposed to use Eurogrids to optimize individual target functions at the nodal level. On the second - the use of the Q-learning algorithm with reinforcement for adaptation to changes in the environment using the forecast of the reward for the maximum gain in the implementation of the target management functions, due to the representation of the resource management process as a Markov process. Using this approach can provide effective network management by adapting to changes in the environment and considering different objective functions.
References
Бєляков Р. О., Фесенко О. Д. Модель мобільності наземної комунікаційної мережі спеціального призначення. Computer-Integrated Technologies: Education, Science, Production. 2023. № 51. С. 130–138. DOI: 10.36910/6775-2524-0560-2023-51-17.
Романюк В. А. Концепція ієрархічної побудови інтелектуальних систем управління тактичними радіомережами класу MANET. НВЧ-техніка та телекомунікаційні технології : Збірник тез доповідей Міжнар. кримська конф. КриМіKo. Севастополь, 2012. С. 265.
Романюк В. А. Цільові функції оперативного управління тактичними радіомережами. Збірник наукових праць ВІТІ НТУУ «КПІ». 2012. № 1. С. 109–117.
Романюк В. А. Архітектура системи оперативного управління тактичними радіомережами. Збірник наукових праць ВІТІ НТУУ «КПІ». 2009. № 3. С. 70–76.
Deep Reinforcement Learning Aided Packet-Routing for Aeronautical Ad-Hoc Networks Formed by Passenger Planes / D. Liu et al. IEEE Transactions on Vehicular Technology. 2021. Vol. 70, no. 5. P. 5166–5171. DOI: 10.1109/tvt.2021.3074015.
Enhancing Vehicular Ad Hoc Networks’ Dynamic Behavior by Integrating Game Theory and Machine Learning Techniques for Reliable and Stable Routing / N. Phull et al. Security and Communication Networks. 2022. Vol. 2022. P. 1–11. DOI: 10.1155/2022/4108231.
Implementation of Mobility Management Methods for MANET / J. Hosek et al. International Journal of Advances in Telecommunications, Electrotechnics, Signals and Systems. 2012. Vol. 1, no. 2-3. DOI: 10.11601/ijates.v1i2-3.39.
Model of data flow control subsystem of the MANET class mobile radio network control system / Yu. Kramska et al. Scientific Journal of TNTU. 2022. Vol. 107. No 3. P. 51–59.
Romaniuk V. A., Bieliakov R. О. Objective control functions of FANET communication nodes of land-air network. Computer-Integrated Technologies: Education, Science, Production. 2023. No. 50. P. 125–130. DOI: 10.36910/6775-2524-0560-2023-50-19.
The Hierarchical Model Of Interaction Between Intelligent Agents In The Manet Control Systems / O. Y. Sova et al. Information and Telecommunication Sciences. 2015. No. 1. P. 21–28. DOI: 10.20535/2411-2976.12015.21-28.
Жебка В.В. Дослідження методів машинного навчання та їх застосування для прогнозування відтоку користувачів телекомунікаційних послуг. Зв’язок. 2020. № 4. DOI: 10.31673/2412-9070.2020.042231.