2019/11/27 by Taha Eghtesad, Eghtesad, Taha, Yevgeniy Vorobeychik +3
Computer Science · Engineering · #Artificial Intelligence (cs.AI) #Computer Science and Game Theory (cs.GT) #Cryptography and Security (cs.CR) #FOS: Computer and information sciences #Information and Cyber Security #Network Security and Intrusion Detection #Smart Grid Security and Resilience
paper · pdf · doi:10.48550/arxiv.1911.11972
openalex publication_date 2019/11/27 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
Moving target defense (MTD) is a proactive defense approach that aims to\nthwart attacks by continuously changing the attack surface of a system (e.g.,\nchanging host or network configurations), thereby increasing the adversary's\nuncertainty and attack cost. To maximize the impact of MTD, a defender must\nstrategically choose when and what changes to make, taking into account both\nthe characteristics of its system as well as the adversary's observed\nactivities. Finding an optimal strategy for MTD presents a significant\nchallenge, especially when facing a resourceful and determined adversary who\nmay respond to the defender's actions. In this paper, we propose a multi-agent\npartially-observable Markov Decision Process model of MTD and formulate a\ntwo-player general-sum game between the adversary and the defender. Based on an\nestablished model of adaptive MTD, we propose a multi-agent reinforcement\nlearning framework based on the double oracle algorithm to solve the game. In\nthe experiments, we show the effectiveness of our framework in finding optimal\npolicies.\n