ERDQN Data Scheduling Algorithm For Low-latency Transmission

Low latency and high stability network transmission is one of the key issues in the development of Internet and Io T.With more and more devices accessing the network,the requirements of different devices for network latency are becoming more and more complex and new applications such as in-vehicle networks,telemedicine,industrial control,etc.have strict requirements for network transmission stability and low latency,so improving network transmission stability and reducing network latency have become hot issues in the current network transmission research.For the problem that the sender of network transmission cannot satisfy the priority and time constraint(Deadline)of different types of data packets,this paper proposes a priority-based G/G/1 network model,which firstly calculates the survival time of data packets by using the creation time and current time of data packets,secondly gets the remaining survival time of data packets according to the survival time and deadline of data packets,and finally calculates the priority of data packets according to the priority and remaining survival time of data packets.Finally,the priority of the data group is calculated based on the priority of the data group and the remaining survival time,and the priority of the data group is used as the criterion to decide the order of the data group into the queue waiting queue.The algorithm satisfies the priority and time constraints of data packets and reduces the queuing delay of data packets in the sending process.To address the problems of low average transmission rate and large transmission rate variation caused by the mismatch between transmission rate and network bandwidth in network transmission,this paper proposes an optimized Experience Replay DQN(ERDQN)algorithm,which first uses the current network link state and possible network actions as feature vectors,and inputs them into the algorithm to predict the next The algorithm is used to predict the network transmission state at the next moment and is given different reward factors.The algorithm solves the problem that the traditional congestion control algorithm has a large adjustment range for the transmission rate,and uses the current network state to adjust the network transmission parameters autonomously,which improves the average transmission rate and stability of network transmission and reduces the transmission delay in the network delay.The experimental results show that the priority-based G/G/1 network model satisfies the priority and time constraints of data packets and reduces the queuing delay in the sending process.The ERDQN-based congestion control algorithm improves the average rate and stability of network transmission and reduces the transmission delay during transmission.In summary,the low-latency ERDQN data scheduling algorithm combining the two models can provide stable transmission with low latency and high reliability.