چكيده به لاتين
With the advancement of wireless technologies and the advent of fifth generation networks, data traffic has greatly increased. In general, in each generation, a lot of effort has been made to provide new services with higher quality to users by using new technologies. On the other hand, one of the needs of this network is more throughput. In dual-connection ferrule connections using the ability to receive data simultaneously from a macro base station and a small base station, a higher throughput can be achieved than in the case where this feature is not used. To this end, part of the traffic flow (to the user) split from the macro base station to the small base station so that both can play a role in sending and receiving data simultaneously by the user. However, due to the variability of channel conditions in air interfaces (and its effect on the delay of transmitter queues) and the possibility of congestion in the backlinks, there is a possibility of differences in the delay of some packets that are sent from two different routes. Given the importance of receiving packets in a specific order in most applications (such as TCP-based applications), flow splitting is important to minimize latency differences in macro and small base stations queues to expedite packet delivery to the application layer on the user side. Therefore, in this algorithm, in order to optimize the branching of traffic flow in dual connections based on integrated access and backhaul technology, for heterogeneous wireless networks, which can be provided in the absence of the statistical model of the system (such as: channel, traffic arrival pattern, and backhaul quality). act Also, unlike the previous works by reducing the delay, the proposed algorithm also takes into account the difference of opinion and keeps it at a threshold level, and gives the degree of freedom and more details to the system status and traffic splitting decision, which according to the results The simulation resulted in a 10% improvement over existing work independent of the model.
