Analysis Of Distributed Intelligent Agent Model For Qos Dynamic Scheme In Gsm/Gprs Network

In this paper we study dynamic Quality of Service scheme in GSM/GPRS wireless network. A load balancing architecture constructed by distributed intelligent agent has been presented to support real time or burst data services. Fuzzy neural network was employed to predict GPRS traffic by learning examples. Meanwhile, we have presented a traffic estimation algorithm and a simple decision mechanism to deal with special applications such as burst data transmission. The simulation shows that distributed intelligent agent architecture could significantly reduce packet delay, route cost and relieve GPRS bottleneck.