Adaptive wavelet neural network controller for AQM router in TCP network

This paper presents a wavelet neural network (WNN) controller based on adaptive learning rates (ALRs) method, for active queue management(AQM) in end-to-end TCP network. The AQM is important to regulate the queue length and short round trip time in TCP network. The WNN controller using ALRs adaptively controls the dropping probability of the TCP network. Also the proposed controller is intelligently trained by GD algorithm. The parameters of WNN are tuned by ALRs method. We apply Lyapunov theorem to verify the stability of WNN controller using ALRs. The simulation results show that the performance of WNN controller using ALRs is superior to that of WNN controller using fixed learning rates.