蚁群链多层博弈网络监护信息融合特征优选  

Ant Chain Multiplayer Gaming Network Monitoring Information Fusion and Feature Optimization Selection

作　　者：王淑敏[1] 孔小婧[1] 

机构地区：[1]许昌职业技术学院信息工程系,河南许昌461000

出　　处：《科技通报》2015年第2期146-148,共3页Bulletin of Science and Technology

摘　　要：在网络安全预测监护模型设计中,需要对网络安全监护信息进行数据融合和特征优选,以提高对变异特征的识别能力。传统方法中,采用蚁群算法进行监护信息特征优化融合进化和链路模型设计,算法无法实现相邻簇头之间的信息素融合,特征优化效果不好。针对这一问题,提出蚁群链运动多层博弈的网络监护信息融合特征优选算法,构建多层博弈网络监护数据样本驱动空间权矩阵模型,引入粗糙集理论,对蚁群引导的粗糙集前馈补偿网络进行动态博弈,实现网络安全监护数据的预测控制目标函数最佳寻优。构建多层博弈网络监护系统模型,得到蚁群链运动的监护信息数据状态跟踪模型,实现网络安全监护信息的融合特征优选改进。仿真实验表明,该算法能有效提高对异常信息的监护和检测能力,有较高的特征优选品质,展示了本文算法在对网络安全监护中的优越性能。In the network security forecast monitoring model design, it needs for data fusion and feature selection of network security monitoring information, in order to improve the recognition ability of variability. In the traditional method, the ant colony algorithm for monitoring information optimization fusion evolution and link model design, the algorithm cannot be achieved between the adjacent cluster heads of pheromone fusion, and feature optimization effect is not good. Aiming at this problem, the network monitoring information fusion and feature selection algorithm is proposed based on ant colony chain motion multi game, a multilayer monitoring data of game network sample driven spatial weight matrix model is constructed,the rough set theory is introduced to guide the ant colony, rough set feed forward compensation network is a dynamic game,it is used to forecast the network security monitoring data and control the objective function optimization. Multi game network monitoring system model is constructed, the information monitoring data state ant chain motion tracking model is obtained, and the integration of feature selection of network security monitoring information is completed. Simulation results show that, the algorithm can effectively improve the abnormal information monitoring and detection capability, feature selection of higher quality is improved, it shows the algorithm is superior in performance of the network security monitoring.

关 键 词：蚁群算法 特征优选 信息融合 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]