基于遗传算法进化业务冲突检测规则的研究  

Evolving the Detection Rule of Feature Interaction by Genetic Algorithm

作　　者：魏薇[1] 杨放春[1] 

机构地区：[1]北京邮电大学网络与交换国家重点实验室,北京100876

出　　处：《电子学报》2007年第4期634-639,共6页Acta Electronica Sinica

基　　金：国家杰出青年科学基金(No.60125101);国家973基础理论研究计划基金(No.2003CB314806)

摘　　要：完备和正确的检测规则是业务冲突管理器在线检测冲突时提高检测率的关键.该文借鉴遗传算法的随机搜索能力提出一种冲突检测规则进化算法,通过加入已知冲突信息的指导和以概率递减选择新业务参与变异的方式提高系统进化速度,对检测规则集进行优化降低系统时间和空间复杂度.实验证明此算法提高了系统检测率.The completeness and correctness of detection rule set are key points to enhance the real-time detection rate of the feature interaction manager （FIM）.Based on the random search capability of the genetic algorithm,an evolution algorithm of detection rule set is proposed.In the algorithm,the known interactions are added and the new services are selected to mutate by the degressive probability, which improve the evolution speed. Optimizing the detection rule set can reduce the time and space complexity. The experiment results show that the algorithm advances the performance of the FIM.

关 键 词：下一代网络 业务冲突 检测规则 遗传算法 检测率 

分 类 号：TN915.5[电子电信—通信与信息系统]