Malicious program patch control method based on improved DDPV algorithm

Abstract

In the current context of information data sharing, with the rapid development of communication technology and intelligence, malicious program patch control technology has been widely applied. To address the issues of malicious program propagation control, this study uses the deep deterministic gradient algorithm to design a malicious program patch control method. On the basis of analyzing the propagation mechanism of malicious programs, intrusion detection systems are used to model the malicious programs. The rehabilitation model for susceptible infected individuals is applied to describe the process of malware transmission and construct a composite malware patch propagation model. A malicious program patch control method based on the dual deep Q-network algorithm is designed by introducing a composite malicious program patch propagation model. The dual deep Q-network algorithm could achieve network equilibrium in 45 time steps. Under the attack of malicious programs with different hit rates, the peak proportion of susceptible devices reached 0.07, 0.02, and 0.286, respectively. The number of devices infected by high hit rate malicious programs was 2.81 times that of devices infected by low hit rate malicious programs. In dynamic network environments, the DDPV method showed good adaptability, could effectively control the propagation of malicious programs under different dynamic conditions, and maintained high network gains and patch success rates. Therefore, adopting the designed malicious program patch control method can effectively suppress the spread of malicious programs by quickly identifying and sending patches, providing strong support for building a secure network environment.