The Strategy Of Load Redistribution In Complex Networks And Its Application

With the rapidly development of network information technology especially for Internet, we have already stepped into the network society, while human beings are on the threshold of a new era in complex networks. However, complex networks are same to a double-edged sword since the network society brought not only great convenience but also some negative effects to people’s daily life. The cascading failure of complex networks is one of the drawbacks brought by the network society, which has caused great losses to our daily life. In this paper we major study the load redistribution strategy of complex networks to improve the robustness of the network. Simultaneously, we present several load redistribution strategies which are different from the existing ones to prevent the catastrophes of cascading failures. The effectiveness of the proposed strategy is verified on the artificial networks and the power network. The main contents and contributions of this dissertation are as follows:Firstly, we propose a local load redistribution strategy based on the residual capacity of the nodes. In previous studies, when the load of failure node is distributed to the candidate nodes, the load of candidate nodes can be allocated from the failure node according to their own load. This principle of distribution cannot meet the real situation perfectly. In contrast to the load redistribution strategy based on the real time load at the nodes, we propose a local load redistribution strategy based on the residual capacity of the nodes. When a node fails, the candidate nodes at the local area are in accordance with the residual capacity to carry out the load from the failure node. The effectiveness of the proposed strategy is verified by the simulation on the artificial networks and the power network.Secondly, we propose a local load redistribution strategy based on the maximum residual capacity of nodes. When a node fails, we search for the nodes in which region is centered on the failure node. Some of the nodes which have the maximum residual capacity in this region are identified as nodes that can be allocated, and the number of the candidate nodes equals to the degree of the failure node. It can eliminate the node whose load is almost same as its capacity to improve the network robustness. The effectiveness of the proposed strategy is verified by the simulation on the artificial networks and the power network compared with the load redistribution based on the residual capacity of nodes.Thirdly, we propose a global load redistribution strategy based on maximum residual capacity of nodes. When a node fails, we can search for the nodes in global networks according to their residual capacity. Furthermore, we select the nodes with larger residual capacity as the candidate nodes. This method not only eliminates the node whose load is almost same as capacity to improve the network robustness, but also reduce the number of candidate nodes to decrease the cost of distribution. The effectiveness of the proposed strategy is verified by the simulation on the artificial networks and the power network.