Study on the Application of Dynamic Continuous Trust Assessment Model Based on Clustering Mechanism in Network Security Optimization

Due to the rapid development of network technology and the increasing size of networks, the security of networks in today's society is facing great challenges. The study of network security is based on the study of network security frameworks, connections, and models to improve network security so that the network becomes controllable, manageable, and survivable. In this paper, we propose a dynamic continuous trust assessment model based on the clustering mechanism and apply this model to the network security model optimization problem. The specific conclusions are as follows: (1) The problems in previous clustering algorithms are analyzed, and corresponding solutions are proposed for the problems. Then, the particle swarm algorithm is briefly introduced, the inertia weight coefficients in the particle swarm algorithm are improved, and the dormant mechanism of neighboring node groups is introduced. (2) A new algorithm for calculating direct trust value is proposed. (3) A method for evaluating the overall trust value level of the network is proposed. The method first selects the N nodes with the highest number of interactions and then evaluates the overall trust level of the network in which these N nodes are located by using the associative memory capability of the clustering algorithm. (4) The experimental analysis shows that the randomly generated rule set in the model of this paper tends to evolve to a stable state after 60~80 rounds of operation, with a higher success rate, and at 180~200 rounds of operation, the rule set evolves to a stable state again, with the success rate still improved; after 240 rounds of operation, the network environment is restored to the initial state, and although the roles of some of the nodes are changed, the model still evolves to a high level again. Level. The dynamic continuous trust evaluation model proposed in this paper can be effective in the network security optimization, but the practical application of the model should be done further research.