Research On Attack And Repair Methods Of Network Topology Based On Evolutionary Search

In the vulnerability analysis of complex networks,most research methods are purely based on the topology of the network.What's more,most of them choose the largest connected component to be the vulnerability evaluation indicators.In practical applications,this kind of analysis method is difficult to reflect the non-topological properties and the functional properties of the network.In fact,for networks with traffic transmission as the main function,people pay more attention to the changes in traffic that can be transmitted between nodes.In response to the above problems,this article has carried out related research and exploration,and the main work is as follows:(1)Aiming at the traffic transmission network,this paper analyzes its transmission efficiency changes in multiple attack modes,and designs the network transportation efficiency function,which is specifically used to describe the degree of decline in the effectiveness of the network after it is damaged.This indicator is based on the development of betweenness,and comprehensively considers the topological structure of the network and the non-topological properties of the nodes.Experiments show that this indicator can represent the function of the network better than the traditional one.(2)After determining the evaluation index of vulnerability,we uses this indicator to describe the effect of network attack and defense in this article,and defines the cost of network attacks and repairs.After modeling and simulating the network attack problem,the network vulnerability problem is transformed into a multi-objective optimization problem between attack repair effect and cost.Aiming at the multiobjective problem,this paper uses the evolutionary algorithm to solve it.In the experiment of Shanghai rail transit network,we have obtained some efficient attack schemes and repair schemes.(3)When applying multi-objective evolutionary algorithms,we designed an improved evolutionary algorithm reproduction operator based on KNN-graph and used it to improve existing evolutionary algorithms.Tests on standard data sets show that the improved algorithm has better diversity and convergence than the original algorithm.Finally,on the Shanghai rail transit network,we applied the improved algorithm to solve the network attack and repair problem.Finally,experiments show that our algorithm can design a better attack plan and repair plan.