Research On Security Analysis Algorithm Based On Bayesian Network Attack Graph

At present,with the rapid development of computer technology and data communication technology,the application of services through the Internet is becoming more and more extensive.It is a hotspot in the field of network security to find a reasonable and practical method for network security analysis.The network attack graph model describes the dependencies between vulnerabilities in the target network and is a tool for analyzing network vulnerabilities.Based on the attack graph generation method,this paper uses the attack graph technique,combined with the security event and the vulnerability grading standard to analyze the vulnerability of the vulnerability node quantitatively.The results of quantitative analysis can help the system administrator to better understand the current network system node security status,so as to further implement the protection strategy to provide reference,reduce the network system security risks.The main work of the article includes:(1)Based on the attack graph generation tool and the vulnerability information in the network system to generate a directed acyclic structure of the attack graph.The vulnerability assessment system is used to evaluate the probability that the nodes are attacked in the attack graph,and then generate the attack graph based on the Bayesian network with directed acyclic graphs.Then,in the presence of the node dependency condition,the prior probability of the node in the Bayesian network attack graph is calculated and the degree of threat is evaluated.Combined with the security event information,change the node state value in the attack graph,and use the Bayesian network posterior probability calculation method to update the nodes in the attack graph to exploit the probability of using the nodes to accurately evaluate the nodes in the attack graph.(2)Aiming at the problem that the complexity of posterior probability calculation is exponentially increasing due to the increase of attribute nodes,a probabilistic method based on group tree propagation algorithm is proposed.Based on the Bayesian network attack map,the adaptive genetic algorithm is used to solve the optimal elimination order of triangulation,and the selection operator of the adaptive genetic algorithm is improved to improve the global maximum of the algorithm.The optimal search ability and the convergence speed.The experimental results show that the adaptive genetic algorithm using the improved selection operator can obtain better order of elimination by simple genetic algorithm and adaptive genetic algorithm,and further reduce the tree weight index which reflects the time complexity of clique tree reasoning.