关键词: 攻击路径预测, 改进蚁群算法, 防护策略, 概率攻击图

Abstract: The multi-step feature of network attacks increases the difficulty of predicting attack paths and makes it difficult to provide effective security against attacks. Traditional solutions take a high cost to repair a large number of network vulnerabilities. In order to solve the above problems, this paper studies the protection of network attacks and proposes a Hardening Measure Selection Mode based on an Improved Ant Colony Optimization（HMSMIACO）. The model consists of three parts. Firstly, based on the existing attack graph, a Bayesian belief network, which can describe the relationship between multi-step atomic attacks, is used to construct a probabilistic attack graph for assessing network security risks. Secondly, a path prediction algorithm that can simulate the attacker’s decision-making process is proposed based on quantitative indicators of defense costs and benefits. Thirdly, considering the selection of protection strategy is an NP-hard problem, this paper chooses an improved ant colony algorithm which is suitable for medium-scale network environment to solve the problem and obtains the optimal protection strategy set in the network environment. Finally, the experiments show that HMSMIACO is feasible and effective in reducing the network security risk.

Key words: attack path prediction, improved ant colony optimization, hardening measure, probabilistic attack graph