Cooperation among malicious agents: a general quantitative congestion game framework

Malicious behaviors and cooperation have been well studied separately. However, rare systematic study has been conducted on the combination of them: malicious cooperation. In this paper, a general quantitative utility function of malicious cooperation is firstly formulated in a congestion game framework. Both objective and subjective factors are incorporated (e.g., malicious social networks and moral degrees). Then, Nash equilibrium and the condition of malicious cooperation are given theoretically. Meanwhile, we show empirically that malicious cooperation may even improve system performance (i.e., catfish effect).