Annotation: This paper is devoted to an approach to countering network attacks based on network reconfiguration to exclude the possibility of successful completion of the attack. To implement the approach, it is proposed to use the mechanism of recommender systems that provides both generation of possible network topologies and their ranking. The proposed intelligent recommender system is based on a reinforcement learning algorithm based on the actor – critic model. Experimental studies have confirmed the effectiveness of the developed system

Keywords: reinforcement learning, DDoS, network attack, recommender system