Research On Adaptive RED Algorithm In Network Congestion Control

The network is affecting the development of the whole world,whether the Internet or Internet,and artificial intelligence,at this stage the fiery cloud computing and big data,the development of cutting-edge technology in the field of all cannot do without a good network service platform,to enable the network to do a solid backing era of economic development,in large data network age must have a strong ability of data transmission,and network congestion control is the key point to improve the capability of network transmission,so how to avoid the congestion of data has become a current topic worthy of study direction.This paper mainly focuses on the improvement of RED(Random Early Detection)algorithm in queue management algorithm.First,aiming at the phenomenon that the RED algorithm is prone to global synchronization,this paper proposes an adaptive RED improved algorithm based on fuzzy control.Secondly,in view of the parameter sensitivity of RED algorithm,a RED algorithm with parameter adaptive dynamic adjustment function is proposed in this paper.The main contents of this paper are as follows:1.An improved parameter adaptive RED algorithm based on fuzzy control-FARED is proposed.Based on the RED algorithm,we optimize the three variables of RED algorithm.The three variables are speed,average queue length and the current queue length.For each parameter variable,the fuzzy controller is optimized,and every fuzzy controller is quantized.Finally,the fuzzy controller is integrated.By adding parameters ? and ? matching the weights of the controllers,an accurate packet loss rate is finally output.Because the fuzzy controller is mainly optimized for the parameters of different variables,it has the mechanism of adaptive parameter adjustment to improve the performance of the network.2.A RED algorithm-DARED algorithm based on adaptive dynamic adjustment of parameters is proposed.The improved algorithm is mainly divided into two steps,one is to discard probability function of the RED algorithm is a linear,or loss of the traditional RED algorithm and semi Cauchy distribution function rate function in nonlinear processing by the S,the parameters of traditional RED algorithm are the defects of static setting,use range the captain and the average queue length parameter adaptive dynamic adjustment strategy to improve the maximum packet loss rate,to improve the performance of DARED algorithm.3.The two algorithms are simulated in NS2.Simulation results show that The FARED algorithm has better improvement in the packet loss rate,average queue length,delay jitter and throughput.Moreover,the global synchronization phenomenon of RED algorithm has also been improved.Besides the slight increase of delay and delay jitter,the DARED algorithm has greatly improved in the three performances of packet loss rate,throughput and average queue length,and has improved the sensitivity of RED algorithm.