Research Of Dynamic Routing Algorithm Based On Local Information In Scale-free Network

Since the small-world phenomenon and scale-free feature of networks are proposed,the study of complex networks has been developing rapidly.Complex networks have penetrated into different disciplines and engineering fields,providing new theories and research methods for people to study real networks.One of the most important functions of networks is to transmit the objects carried by the network,such as the transmission of data packets on the Internet and traffic flow in traffic systems.Studying the interaction between network topology and traffic dynamics by exploiting the theory of network science,it can help people to understand and solve the problems in real networks.Nowadays the scale of the networks is increasing rapidly,and the emergence of the big-data technology will inevitably increase the data flow in the network.In this case,the congestion often occurs,which restricts the transmission performance of the network.In order to achieve the transmission requirements of modern networks,it is a hot research topic to alleviate the congestion of networks and improve the transmission capability.In real life,most networks exhibit scale-free features,and the global information of the network is hard to be obtained due to the increasing scale of the network.So,it is practical significant to improve network transmission capability and alleviate the congestion by exploiting the local information on scale-free networks.In addition,there is not only scale-free features,but also degree correlations between nodes in the networks of real life.In order to coincide with the nature of the real network topology,it is necessary to study the transmission capability of the degreeassociated network.Therefore,the main contents of this paper are as follows:(1)A dynamic routing algorithm based on the node degree and the length of queue of node is proposed.This algorithm only need to search for the information of neighbor nodes and next nearest neighbor nodes,and thus it can avoid traversing the topology information of the entire network.At the same time,two adjustable parameters ? and ? are introduced,by adjusting these two parameters,the transmission capacity of the network will be changed,and the optimal parameter combination of the algorithm is found.At the same time,we also study the packet loss rate of routing algorithm.The simulation results show that the algorithm can not only effectively improve the network transmission capacity and alleviate the network congestion,but also reduce the packet loss rate in the network greatly.(2)The influence of the network with degree correlation on transmission capacity.The impact of degree correlations on the transmission capacity is studied.Considering that there are both scale-free features and degree-associativity between nodes in real networks,we first construct scale-free network model with degreeassociativity,then the impact of degree correlations on transmission capacity is studied.The simulation results show that the degree correlation has an important influence on the transmission capacity of the network.The greater the degree correlation of network is,the lower the transmission capacity is.(3)Verification and analysis of empirical data.In order to verify the performance of the network transmission capacity based on the routing algorithm we proposed on actual network,we obtain two real network data: Internet AS data and a university Email system data,which their degree correlation coefficient is-0.2 and 0.08 respectively.We compare 3 different local routing algorithms,experimental results show that the dynamic local routing algorithm we proposed improve the transmission capacity more than 2 times than others.