Intelligent Parameter Tuning using Deep Q-network in Adaptive Queue Management Systems

Network traffic has risen in recent years to the point that it is obviously and astonishingly 2020, with increase predicted double following days. Up 23 Teraa bit every month an incredible amount. The Active Queue Management (AQM) algorithm one of most significant study areas network congestion control; nevertheless, new self-learning management algorithms are needed on nodes cope huge quantity minimize queuing latency, used reinforcement learning for automatic adaptive parameter AQM effective management, present a novel focuses deep deal latency trade-off between delay throughput; choose Deep Q-Network (DQN) as foundation our scheme equate Random Early Detection (RED) Results based simulation (NS3) suggest DQN good better results were obtained from RED, where difference reached drop rate 2%, this percentage considered good, addition throughput packet transfer 3% proposed algorithm. Index Terms— Adaptive Management, congestion, Traffic Q-Network, Reinforcement Learning.