Security games with ambiguous beliefs of agents

Abstract

Currently the Dempster- Shafer based algorithm and Uniform Random Probability based algorithm are the preferred methods of resolving security games, in which security forces are able to identify attackers and need only to determine their strategies. However this model is inefficient in situations where resources are limited and both the identity of the attackers and their strategies are ambiguous. The intent of this study is to find a more effective algorithm to guide the security forces in choosing which outside forces with which to cooperate given both ambiguities. We designed an experiment where security forces were compelled to engage with outside forces in order to maximize protection of their targets. We introduced two important notions: the behavior of each agent in target protection and the tolerance threshold in the target protection process. From these, we proposed an algorithm that was applied by each security force to determine the best outside force(s) with which to cooperate. Our results show that our proposed algorithm is safer than the Dempster- Shafer based and Uniform Random Probability based algorithm.