中文摘要：

在对多服务器环境下的网络最脆弱攻击区进行检测的过程中，由于随机性较大，容易产生很多冗余脆弱攻击区，造成干扰。导致传统核Fisher判别分析，无法获取最优网络最脆弱攻击区特征．不能有效实现多服务器环境下的网络最脆弱攻击区检测等问题，提出一种利用AFSA—SVM的多服务环境下的网络最脆弱攻击区检测方法，对多服务器环境下网络状态信息进行采集，提取网络状态特征。将多服务器环境下网络特征子集编码成人工鱼的位置，通过仿真鱼群的觅食、聚群及追尾行为获取最优特征子集，将最优特征子集作为SVM分类算法的输入，使多服务器环境下的网络最脆弱攻击区检测问题变成了对多个二分类问题，进行求解，把得到的结果结合在一起，获取多服务器下的网络最脆弱攻击区检测结果。仿真结果表明，所提方法具有很高的检测精度及检测效率。

英文摘要：

On multiple servers of network is the most fragile zone of environment in the process of testing, due to the randomness, prone to a lot of redundant fragile zone, causing interference, lead to the traditional kernel Fisher discriminant analysis, unable to get the optimal network the most vulnerable to attack area characteristics, cannot effectively realizing the server of network is the most fragile zone of environment detection problems, put forward a kind of AFSA - SVM more service of network is the most fragile zone of environment detection method, the network status in multi-server environment information collection, to extract the characteristics of network status. Multi-server environment network feature subset codes into the position of the artificial fish, the fish of foraging behavior, cluster and fundation for obtaining the optimal attributes subset and optimal feature subset as input of the SVM classification algo- rithm, make more server of network is the most fragile zone of environment detection problem into multiple binary classification problems, to solve, combine the results together, under multiple server access network is the most fragile zone test results. The simulation results show that the proposed method has high detection accuracy and efficiency.