The strategy research and method realization for the computer network flow control
Jian-qing Du · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · 2013
FARIMA (p, d, q) model fitting the real network traffic and the input of noise is viewed as the non-controllable input in this paper. The differential operator embodies the long and short correlation of self-similar correlation. However, the order is too high because it’s difficult to control, this paper uses Pade reduced-order algorithm to reduce order. And then the reduced expression is converted into noise terms with the use of minimum variance control to eliminate the uncontrollable input caused by the impact of the system output error, making the system output (queue length) stable in the vicinity of thresholds. Finally, the feasibility and stability of the system is applicability on the minimum variance control algorithm through analyzing the theory that is self-similar network traffic control, but there are some differences in the theory.