Inspection Games

Abstract

AbstractDresher began the so‐called inspection game as a study on the treaty of arms reduction in 1962. After him, his work was extended to several branches of research fields. The first branch is inspection games for inspectors to deter illegal behaviors of violators for the compliance of treaties or laws, taking account of violators' intention, for example, the nuclear inspection by the IAEA to prevent the spread of nuclear weapons and technology. The second branch is smuggling games, where Customs tries to interdict the drug smuggling by a smuggler. As the inspection game, we could regard network interdiction games, ambush games, and infiltration games. In this article, we mention those models of the inspection game.First, we review the history of researches on the inspection game. Second, we describe a basic model of the inspection game with multiple stages, which is often taken in many past researches. We modify an assumption of the basic model concerning with acquisition of information about players' opponents and consider several versions of the inspection game: a multistage game with complete information, a one‐shot game with no information, and a multistage game with incomplete information. While outlining the modeling and the formulation of each model, we describe useful methodologies to solve each game, such as difference equation, dynamic programming, and the concept of Bayesian–Nash equilibrium.