无标度网络中的拥塞转变动态分析

Dynamics of Jamming Transitions in Scale-Free Networks

把局部流量信息与最短路径路由策略相结合,提出了一种具有感知流量信息的路由策略算法.在该算法中,存在一个调节最短等待时间和最短传输路径之间权重的控制参数,通过调节这个控制参数可以使网络的传输能力达到最优.在具有不同聚类系数的无标度网络模型中进行仿真,仿真结果表明,拥塞转变被两种不同的相变曲线所描述,并且网络容量的大小取决于网络结构的基本属性和路由策略.与最短路径算法相比,采用该路由算法无论无标度网络的聚类系数如何,网络的吞吐量均得到较大提高,但就该路由算法本身而言,吞吐量随着聚类系数的增加而减小.

Incorporating the local traffic information into the shortest path, a routing strategy that is aware of traffic information is proposed, where a controlled parameter is used to balance the weight of waiting time and the shortest path. Adjusting the controlled parameter, the transmittability of network can be optimized. A simulation is performed in a model of scale-free network with different clustering coefficient, and the results reveal that the jamming transition is characterized by two quite distinct phase transitions and that the network capacity depends on the basic attributes of network architecture and routing strategy. Compared to the shortest path, the network throughput by the strategy proposed is all enhanced greatly no matter what the clustering coefficient is in scale-free network, while, the throughput decreases with increasing clustering coefficient in the routing strategy itself.