Ad Hoc网络中基于概率预测的拥塞控制算法

Congestion control algorithm based on probabilistic prediction in Ad Hoc networks

为缓解网络拥塞,提高网络服务质量,提出一种基于概率预测的拥塞控制算法(PPCC)。处理已有数据来构建贝叶斯网络,找出Ad Hoc网络中表示拥塞状态参数之间的关系;改进贝叶斯网络的推理过程,更好地利用所获网络实时参数预测拥塞状态的概率;根据预测得到的概率自适应地调整网络参数,有效地控制网络拥塞,一定程度上提高网络服务质量。仿真结果表明,PPCC算法能够控制网络拥塞,保证网络运行在具有较小时延、较高吞吐量的状态,更好地利用网络资源。

To relieve the network congestion and improve the quality ot＂ service of networks, a congestion control algorithm based on the probabilisfic prediction （PPCC） was proposed. Firstly, the existing data were processed to build the Bayesian network, which identified the relationships between parameters representing congestion states in Ad Hoc networl Secondly, the probability of congestion state was predicted by improving the deduction process of Bayesian network and effectively utilizing the obtained real-time network parameters. Finally, network parameters were adaptively adiusted according to the predicted prohabilities to control the network congestion effectively and im- prove the quality of service of network to some extent. The simulation results show that PPCC can control the network congestion effectively and ensure the networks in good running states with shorter delay and higher thronghput. It makes better use of the networks resources.