AUTHOR=Wang Zhen , Jiang Mengting , Yang Yu , Chen Lili , Ding Hong 

TITLE=Computing Effective Mixed Strategies for Protecting Targets in Large-Scale Critical Infrastructure Networks

JOURNAL=Frontiers in Physics

VOLUME=Volume 9 - 2021

YEAR=2021

URL=https://www.frontiersin.org/journals/physics/articles/10.3389/fphy.2021.805584

DOI=10.3389/fphy.2021.805584

ISSN=2296-424X

ABSTRACT=Most critical infrastructure networks often suffer from malicious attacks, which may result in network failures. Therefore, how to design more robust defense measures to minimize the loss is a great challenge. There are lots of relative studies that develop algorithms to get defense strategies for enhancing the robustness of complex networks when facing attacks. Few of the works base on game theory, but their algorithms cannot solve the defending problem on large-scale networks with full strategy set effectively. In this paper, based on the existing two- person zero-sum game model and the Double Oracle framework, we propose EMSL algorithm which is an approximation algorithm based on greedy to compute effective mixed strategies for protecting large-scale networks. The effectiveness of the mixed defense strategies obtained by our approximation algorithm is proved, as well as the improvement of the mixed strategies to other defense strategies is discussed. Experimental results show that our approximation algorithm can compute the mixed strategies on large-scale networks with full strategy space, and the mixed defense strategies bring the highest utility to defender on different networks when dealing with different attacks.