| Summary: | The existing cyberspace attack and defense method can be regarded as game, but most of the game only involves network information, not include cyberspace's states, attacker's and defender's actions. To address this challenge, this paper proposed a cyberspace attack and defense game model based on reinforcement learning. By establishing two agents, representing the attacker and the defender respectively, defender will select the actions in the cyberspace to obtain defender's optimal reward by reinforcement learning. In order to improve the defense ability of defender, a game model based on reward randomization reinforcement learning is proposed. When the defender takes the defense action, the reward is randomly given and subject to linear distribution, so as to find the better defense policy and improve defense successfully rate. The experimental results show that the game model can effectively simulate the attack and defense state of cyberspace, and the reward randomization reinforcement learning method has a higher defense success rate than existing methods. © 2001 Elsevier Science. All rights reserved.
|
|---|