基于概率分布自适应的化学反应的元启发式优化算法运用于包匹配  

Packet Matching Using Self-adaptive Chemical-reaction-inspired Metaheuristic for Optimization with Probability Distribution

作　　者：王则林[1,2] 吴志健[1] 尹兰[1,3] 邓长寿[4] 

机构地区：[1]武汉大学计算机学院,武汉430072 [2]南通大学计算机科学与技术学院,南通226019 [3]贵州师范大学数字与计算机科学学院,贵州550001 [4]九江学院信息科学与技术学院,九江330005

出　　处：《计算机科学》2014年第5期164-167,185,共5页Computer Science

基　　金：国家自然科学基金(61070008;61364025);教育部人文社科基金(12YJCZH274);贵州省科学技术基金(黔科合J字LKS[2012]37)资助

摘　　要：包匹配是防火墙、路由器等设备中重要的研究焦点,它的速度直接影响着设备的性能。根据当前群的样本信息,把信息熵和直方图的理念引入当前群的信息统计,进而运用这些信息去动态调整化学反应的元启发式算法的相关参数。首次从当前群样本的角度,而不是假设全体样本的分布去分析问题。从实验结果看,其达到了很好的预期效果。基于化学反应的元启发式智能算法由于参数的动态调整,包匹配的规模和性能之间的联系更弱,从而使此智能算法更适合包匹配。Packet matching is the research focus of firewall and router devices.Its speed influences directly device perforrnances.According to the sample informations of population currently,this paper drawed information entropy and histogram into information statistics of population currently,and used these statistics informations to adjust dynamicaly relevant parameters of optimization algorithm of Chemical-reaction-inspired metaheuristic.This paper,for the first time,analyzed problem from the sample of population currently,instead of assuming what is the distribution of all sample.From the experimental results,the algorithm proposed by this paper gets good desired effects.For dynamic adjustment parameters of intelligence algorithm of Chemical-reaction-inspired metaheuristic,the relation of scale and performance of packet matching is more loose.And the intelligent algorithm proposed is more suitable for packet matching.

关 键 词：包匹配 信息熵 直方图 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]