基于自适应神经模糊推理的计算机网络性能评价研究  

Research on performance evaluation of computer network based on adaptive neural-fuzzy inference

作　　者：张晓琴[1] 黄玉清[1] 梁靓[1] 

机构地区：[1]西南科技大学信息工程学院,四川绵阳621010

出　　处：《计算机工程与设计》2009年第22期5100-5102,5123,共4页Computer Engineering and Design

基　　金：国家863高技术研究发展计划基金项目(2006AA887007)

摘　　要：根据带宽、时延、丢包率3个网络关键性能指标,建立了网络性能评价的自适应神经-模糊推理系统。通过对网络不同业务服务质量进行分析,实现了在给定输入负载下对网络性能的判定。仿真结果表明,建立的自适应神经-模糊推理系统能描述网络性能指标和输出的映射规律,能较准确的拟和数据,评价结果符合规律。因此,该方法合理有效,能够为网络信息传输提供决策支持。Based on the bandwidth, delay and packet loss rate, an adaptive neural-fuzzy inference system for network performance evaluation is designed. Through the analysis of different service quality, the judgment of network performance with given input load is realized. The simulation results show that the adaptive neural-fuzzy inference system reflect the mapping rules of network performance metrics and output, moreover fit data accurately, the results are conform to the regular pattern. Therefore, the method is feasible and effective and provide decision support for network information transmission strategy.

关 键 词：网络性能 神经网络 模糊推理 评价方法 服务质量 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]