Network congestion control problem is one of the most key problems in network study. This paper studies different versions of TCP source port algorithm such as Tahoe, Reno, NewReno, SACK and Vegas, and makes simulation research on linear network and dumbbell network environment algorithm by using NS2 network simulator, and finally it concludes that: when there is no competition in the network, it is better to choose Vegas algorithm; when there is competition, choose Reno, NewReno, or Sack algorithm is more advisable.
Network congestion control algorithm based on Actor-Critic reinforcement learning model