交通流局部单向传输对复杂网络拥塞的影响

实施网络局部交通流的单向传输能否有效地提高网络的吞吐量及缓解网络拥塞情况, 对于许多现实复杂系统而言具有重要意义. 通过研究节点的度和介数与节点拥塞情况的关系, 提出了两种用于交通流单向传输控制的流向约束模型. 为检验不同模型下流向约束对网络传输能力的影响, 通过引入一个状态参数H, 利用由稳态到拥塞状态的指标流量相变值来度量网络的吞吐量, 并分别针对3种网络在不同约束模型下进行了仿真. 通过数值模拟得到如下结论: 采用基于度的流向约束模型对ER随机网络和WS小世界网络的数据传输能力几乎没有影响, 网络吞吐量和拥塞情况均没有得到改善, 但该模型在一定程度上缓解了BA无标度网络的拥塞程度; 采用基于介数的流向约束模型有效地提高了ER随机网络和WS小世界网络的吞吐量并在一定程度上缓解了网络拥塞情况, 且较基于度的约束模型进一步改善了BA无标度网络的拥塞程度, 但无论采用哪种约束模型, 均不能有效提高BA无标度网络的吞吐量. 同时可以看出, 对于3种网络, 有意识地对介数较大(拥塞较严重)节点实施流向约束均可以有效地缓解网络的拥塞程度.

Congestion phenomena in complex networks based on local one-way transmission of traffic flow

Whether implementing one-way transmission of local traffic flow can improve the throughout of the network and relieve the congestion situation or not, is of great importance to many real complex systems. We investigate the relationship between the degree and betweenness of the node and the congestion, and propose two traffic flow constraint models used for the one-way transmission control of traffic flow. In order to test the impact of flow constraint on the network transmission capacity under different models, we introduce an order parameter H to measure the throughput of the network by the critical value of phase transition from free state to jammed state, and study the transmission efficiency of three network models upon different flow constraint models. Simulation results show that, the traffic flow constraint model based on degree which has no effect on the data transmission capacity of ER random networks and WS small-world networks, the throughput and the congestion are both not improved, but relieve the congestion of BA scale-free networks to some extent: when choosing the traffic flow constraint model based on betweenness, the throughput of ER random networks and WS small-world networks are both enhanced effectively, and further improve the congestion of BA scale-free networks compared with the model based on degree. No matter use which kind of constraint model, all cannot improve the throughput of BA scale-free networks effectively. However for these three network models, restricting the nodes with large betweenness (or severity jammed) consciously, may relieve the congestion situation of the network effectively.