Optimization Strategies For Improving Network Capacity Based On Complex Network Theory

The discovery of small-world phenomena and the scale-free characteristics promoted the development of studies on complex networks in the end of the 20th century. People started using the theory of complex networks to understand the dynamic process on the real networks and then solve the problems in the network. Congestion is inevitable and important in real complex networks. The congestion on traffic network, communication network or other networks can reduce the network performance and then lead to the network collapse. It will give rise to the inconvenience and trouble of people’s lives. So the study of how to develop the potential utilization of network and improve the network transmission capacity of is very necessary. This paper points out that the packet easily accumulated in the central nodes and the center nodes become congested then will lead to the entire network congestion. Center nodes easily become the bottleneck of network transport efficiency. The network capacity can be significantly influenced by three ways:developing the routing strategy, optimizing network topology and allocating network resources. The starting point of our study is to balance the traffic load on the nodes of network. We assume that the network is infinite and equally distributed. Based on the former study, this paper presents some theoretical study and experiment result in the first two ways, then proposed some strategies to improve the network capacity, and a lot of simulation results are given to verify their efficiency:(1) On the way of optimizing network topology, a new strategy was proposed by removing links in scale-free networks to improve the network capacity. After analyzed HDF and HBF strategies in detail, we think removing links based on the node’s degree can make networks become homogeneous faster. But betweenness can reflect the traffic load of node more correctly, so removing links based on the node’s betweenness can alleviate the center node’s traffic load. So a new strategy that removing a fraction of links in the network according to both the degree and the betweenness of the nodes was proposed.(2) On the way of optimizing network topology, two new strategies were proposed by adding links in scale-free and random networks. After analyzed the problem of LDF and LBF strategies, two improvement strategies were proposed. The two adding link strategies were proposed to add a fraction of shortcuts in the network according to the betweenness and the traffic load of nodes on the shortest path.(3) On the way of developing the routing strategy, we proposed an efficient routing strategy inspired by the efficient routing strategy. We introduce the link traffic load and we think the node and the link in the network both can become congested. In general, the link with heavy traffic load can lead to its end nodes has heavy traffic load and then become congested. The routing cost function of the new routing strategy both considered the node’s betweenness and the link’s betweenness of the network. The routing strategy is based on the node’s and link’s traffic load.