On Active Window Management Algorithm For Uncertain TCP Networks

With the advent of the artificial intelligence era,data plays an increasingly important role in our daily lives.At the same time,to ensure the stability of network services,network congestion has become a very important issue.The goal of congestion control is to optimize the comprehensive indicators such as link utilization and transmission delay.Because TCP networks use store-and-forward for data transmission,router queue management plays a key role.Active window management,as a queue management algorithm,has proven to be an effective congestion control algorithm.From the perspective of control theory,the TCP network can be regarded as a nonlinear dynamic feedback control system.Based on the analysis of the active window management mechanism,this paper studies the active window management algorithm by using the control theory.The main research work and conclusions are as follows:Firstly,for the TCP congestion problem of single route,an adaptive PID controller based on genetic algorithm is designed.Firstly,the small deviation linearization is carried out at the equilibrium point.The linearization model of TCP network is established and the parameters are designed based on the model.Undetermined PID controller.Then the genetic algorithm is used to optimize the three parameters of the PID controller,and a performance index function that comprehensively adjusts the time,rise time,overshoot and steady state error is defined.The search is quickly performed in a given search space,and the performance is obtained.The PID parameter value with the best indicator.The results show that the proposed method can quickly stabilize the router queue near the reference value and effectively solve the network congestion problem.Through comparative analysis,it proves that for the single-routed TCP network,the active window management algorithm is deployed in the router,and the PID designed by the genetic algorithm is combined.The controller can make the network more convergent and robust.Secondly,for the nonlinear TCP network model,considering the uncertainty of network parameters and the influence of unresponsive UDP stream,an adaptive finite-time controller based on non-singular terminal sliding mode is designed.Firstly,the uncertainties caused by uncertain parameters in the network system and the influence of UDP flow are integrated into a total uncertainty.Since the upper bound value of the system uncertainty is difficult or impossible to measure in actual engineering,a self is designed.The adaptive law adjusts the upper bound value of the uncertainty online,and applies a new law to design a sliding mode controller.The Lyapunov stability theory is used to prove that the system state can reach the non-singular terminal sliding surface in a limited time,which proves The equilibrium state of the system converges for a finite time.Simulation analysis under different loads and UDP interference shows that the designed controller has faster convergence and adaptability to uncertain environments.Finally,the main work done in this paper is summarized,and the future research directions are prospected.