中文摘要：

为了有效判断网络数据包是否存在被攻击的可能性,在以往的研究基础上提出了一种新的检测算法DMPS（Detection method based of particle susarm）。首先该算法根据数据包属性的离散度定义了状态检测指标,并利用粒子群优化方法给出了标准差分布的计算流程,以此判断数据包的异常状况。最后,通过OPNET和Matlab进行仿真实验,深入研究了影响该算法的关键因素,同时对比了与其他算法之间的性能状况,结果表明DMPS具有较好的适应性。

英文摘要：

In order to effectively determine the possibility of attacks for network packets, a new detection algorithm DMPS is proposed by previous studies. At first, the state indicators is defined with the discreteness of packet characteristic in this algorithm, and the calculation process of standard deviation distribution is presented to judge the anomaly of packet by Particle Swarm Optimization. Finally, a simulation with OPNET and Matlab was conducted to study the key factors of DMPS. Compared to the perform ance of other algorithm, the results shows that it has better adaptability.