基于QoS的网络规划

Network Plan Based on Quality of Service

对于计算机网络的研究,传统的算法都是在网络完好情况下进行网络规划,而没有考虑网络故障的情况。该文采用混合遗传算法形成满足连通度和跳数约束的初始网络,分析了网络故障下的运行情况。对由故障引起的链路流量的变化,提出了一种启发式算法,该算法对网络进行了局部的改进。该方法在某一链路故障产生后,进行重新路由,寻找饱和链路,通过增加链路以减轻其它链路上的流量负荷,对某些闲置或利用率低的链路进行删除,以减少网络资源的费用。仿真结果验证了该算法的有效性。

In the context of computer networks, most conventional algorithms are designed to network under no-failure conditions, but the troubles of network are not taken into account. This paper presents a hybrid approach of genetic algorithm to form the initial network, which has hops and edge-connected constraints, and then analyses the network behavior after failure. To the change of the flow due to failure, it presents a heuristic algorithm to make local improvement for network. Using the method, it makes a new routing after a link failure to search for saturation links, then to add links to reduce the flow load of the rest links, finally to delete the leisure links or links with lower using rate to reduce the cost of network. Simulation results show the effectiveness of the method.