一种复杂网络通信抗毁节点优化算法仿真

摘要

In the network communication system,interference makes the node load increase suddenly,causing the compulsory transfer of the node load.Shifting the burden of the node to other nodes may cause the overload failure due to the node quantity limitation,resulting in the cascade collapse of the routing nodes,causing the network control paralysis in large area.In order to address the problems mentioned above,a node selection support vector machine (SVM) incremental learning algorithm was put forward in this paper.Through adding the node cascade self learning process,when large-scale paralysis happens among nodes,the algorithm can calculate the optimal node through self -learning,avoid the failed nodes and overcome the paralysis.Theoretical analysis and simulation results show that the anti-destroying ability of the new algorithm in network communication has been improved greatly and the network robustness enhances noticeably.%在网络通信系统中,干扰使得节点负载增大,会造成节点负荷强制转移,将节点负担转嫁给其它节点,节点数量限制会引起过载失效,造成路由节点级联崩溃,引发网络控制大面积瘫痪.传统的通信算法中只是单纯地增加节点数量规避风险,但是节点的增加不能承担海量转嫁负担,不能很好地规避级联崩溃的风险.为了解决上述问题,提出一种节点选择支持向量机增量学习算法,通过加入节点级联自学习过程,在节点发生较大规模瘫痪的情况下,通过自学习计算最优节点,规避失效节点,克服瘫痪.通过仿真结果分析显示,新算法下网络通信中抗毁性有了较大的提高,网络的鲁棒性明显增强.