模糊自调整的加强型随机指数标记算法

A Fuzzy Self-tuning Random Exponential Marking Algorithm Based on Enhanced Price

主动队列管理是网络拥塞控制中一个重要的研究领域。由于网络环境复杂多变,而基于加强型价格的随机指数标记算法(EPREM)参数固定不变,环境适应性不强,故而很难保证服务质量(QoS)。为了克服上述缺点,提出了参数模糊自调整的加强型随机指数标记算法(F-EPREM)。该算法利用模糊控制原理,根据队列误差大小及其变化率在线自动调整算法参数β、γ和ξ以适应网络环境变化。NS2中的仿真结果表明,相对于REM和EPREM算法,F-EPREM加快了队列收敛速度,提高了队列稳定性,有效提高了主动队列管理算法的鲁棒性。

Active queue management （AQM） is a very important research area in congestion control. Random exponential marking based on enhanced price （EPREM） suffers from some problems such as sluggish response to dynamic traffic due to its fixed parameters and can＇t meet the QoS requirement in the changing network. To overcome such shortcomings, an enhanced EPREM algorithm called F-EPREM is proposed, which can adjust the parameters β, γ and ξ according to the queue error and rate error by using the rule of fuzzy control. The simulation results in NS2 platform indicate that F-EPREM can achieve higher convergence speed, smaller queue jitter, and effectively improve the robustness of AQM scheme.