The Simulation Research Of Complex Network Routing Strategy

The research shows that the actual network such as Internet, power grids and transportation network has the typical characteristics of complex networks, for example, small-world and scale-free. With the development of the society, the actual network has expanded sharply, and its load has increased dramatically. Therefore, how to improve the performance of the network and control the network congestion are big issues for the current research community and industry. In order to facilitate carrying out a variety of experiments on the large-scale actual network, people try to abstract it into the corresponding network model and make full use of the complex network simulation platform for the simulation experiment. This paper has given an analysis and improvement to the current typical simulation platform of complex network, thus carrying out an improving approach to a typical local routing strategy. Meanwhile, based on the simulation platform, a study of the effects on different network model caused by improving routing strategies is induced. Specific work can be divided into the following parts:Firstly, this paper analyzes the function of the complex network simulation platform. We found that: 1. In the aspect of network model structure, the simulation platform only provides a BA network model. But there are many kinds of network model in the real world. In order to more fully embody the actual network, this paper base on the construction algorithm of complex network and sets up grid networks, random networks, random improvement networks, WS and NW small-world models in the simulation platform. 2. In the simulation experiment, the simulation platform lacks visual interface. This paper uses MFC to set up a human test interface. Through the interface, researchers can choose to build a network model and set up the required parameters in the process of simulation.Secondly, this paper proposes an improved method according to the traditional local routing policy on the BA network. This method adds the packet queue length of node as the main parameter. It can adjust the processing capacity of nodes and the packet queue length of node by the adjustable parameter ɑ. BA network simulation experiment has shown that the network performance is optimal when the parameter ɑ = 3. What’s more, due to the increase of the dynamic parameters of the packet queue length of node, the routing algorithm can more accurately judge the node congestion situation and determine the forwarding path, therefore, the improved routing strategyon the improvement of network performance is better than the classical local static routing strategy.Thirdly, in order to make the research more generally, we conduct simulation experiments respectively in the random network and WS small-world network to study the effect of new strategy on network performance. The experimental results show that the network performance of random network and WS small world network is optimal when the parameter ɑ = 3. The main reason is that: with the same parameters of simulation environment, BA network has a big difference with the above two networks in topology structure. There exist a few maximum node degree in BA network; however, node degree of random network and WS small-world network distributes themselves more evenly, and has no maximum node degree, which causes the BA network and the two evenly distributed networks to have a big difference in optimal parameter values.