A TCP friendly rate control algorithm based on GRU prediction model
Kuai Yu, Zhaohua Long · 2021
In order to solve the congestion control problem encountered in the transmission of real-time data in the wireless network, combined with the GRU neural network, an optimized TCP-friendly rate control algorithm based on the GRU prediction model is proposed. Predict the sending rate of the sender in the next stage, so that TFRC can adjust the sending rate more timely and accurately. Experiments have proved that the improved algorithm can optimize the congestion caused by blind adjustment of the sending rate to a certain extent while ensuring TCP friendliness Packet loss problem.