Аналіз методів прогнозування змін маршрутів передачі даних в бездротових самоорганізованих мережах
| Jazyk: | angličtina |
|---|---|
| Rok vydání: | 2020 |
| Předmět: |
Artificial neural network
lcsh:Military Science Wireless network Computer science media_common.quotation_subject мобільні радіомережі військове призначення зміна маршрутів передачі даних прогнозування перевантаження штучний інтелект lcsh:U computer.software_genre Field (computer science) Packet loss Traffic optimization Quality (business) Data mining Routing (electronic design automation) computer Data transmission media_common |
| Zdroj: | Збірник наукових праць Харківського національного університету Повітряних Сил, Iss 1(63), Pp 60-67 (2020) |
| ISSN: | 2518-1661 2073-7378 |
| Popis: | The article analyzes the current research in the field of forecasting, which examines the methods of maintaining the quality of MR operation and forecasting the time of congestion of data transmission routes. Traffic queues have been identified as an integral feature of packet-switched networks. The main factors that influence the quality of MR operation are such as the pulse of incoming traffic, the level of bandwidth utilization, sensitivity to delays or packet loss. The following are methods of ensuring the quality of traffic management services: traffic forecasting, feedback, reducing network load. The methods of forecasting changes of data transmission routes, which include intellectual, intuitive, formalized, cause-effect and principles of their operation and functioning, are considered. The implementation of classical methods is exemplified by the method of time prediction based on neural networks, fuzzy neural networks, evolutionary computations, complex control, hybrid fuzzy-stochastic method, traffic balancing, traffic optimization, and detection of temporal anomalies. The advantages and disadvantages that determine their effectiveness under different conditions of application are indicated. On the basis of the analysis the expediency of combining intellectual, cause-effect, formalized and intuitive methods, namely artificial intelligence, evolutionary calculations and parameter management, was determined. Considering the functional features of MR construction and the requirements for the methods considered, the solution is to build methods based on the combination of artificial intelligence, evolutionary calculations and parameter management, which in turn provides the opportunity to develop new methods for predicting changes in wireless data routing in wireless networks. networks. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|