自体集网络入侵检测中的高效寻优算法仿真

发布时间：2018-05-30 01:06

本文选题：网格计算 + 任务调度　；参考：《计算机仿真》2013年08期

【摘要】：延迟容忍网络环境复杂,自体集数据规模较大造成的时空上的巨大消耗而难以处理,导致传统网格入侵匹配的数据分类过程收敛速度慢、局部最优等缺陷。为了提高网格入侵特征匹配效率,提出一种基于重复博弈的自体集网络入侵检测中的高效寻优算法。算法利用了入侵数据特征在空间上的相对集中性,在分类概率计算中,加入一种约束计算过程的博弈因子,在概率匹配过程中,通过数据博弈消除多次概率对比,并加入博弈约束计算最优反应函数,将最优反应函数求解过程看作是一个博弈的过程,计算分类优化达到的极大值。仿真结果表明,改进算法的分类匹配时间复杂度有了一定程度的降低,优化效果较为明显。
[Abstract]:The delay tolerance network environment is complex and the large scale of autologous set data makes it difficult to deal with the huge consumption in time and space which leads to the traditional grid intrusion matching data classification process convergence speed is slow local optimum and other defects. In order to improve the efficiency of mesh intrusion feature matching, an efficient search algorithm for self-set network intrusion detection based on repeated game is proposed. The algorithm takes advantage of the relative centrality of intrusion data features in space. In the calculation of classification probability, a game factor is added in the process of constrained computation. In the process of probability matching, the multiple probability contrast is eliminated by data game. The solution process of the optimal response function is regarded as a game process and the maximum value of the classification optimization is calculated by adding the game constraints to calculate the optimal response function. The simulation results show that the time complexity of the improved algorithm is reduced to a certain extent, and the optimization effect is obvious.
【作者单位】：山西大学商务学院;
【基金】：山西省自然科学基金项目(2010011022－2) 2011年度山西省高校科技开发项目(20111134) 2011年度山西省高等学校教学改革项目(J2011117) 2011年度山西大学商务学院科研基金项目(JG201102)
【分类号】：TP393.08

【相似文献】