Strategic Concealment of Environment Representations in Competitive Games

Abstract: This paper investigates the strategic concealment of map abstractions used by the players in competitive games. We consider a defense scenario in which one player (the Defender) seeks to infer and exploit the abstraction used by the other player (the Attacker). The interaction between the two players is modeled as a Bayesian game: the Defender selects a barrier configuration, i.e., a placement of obstacles that can obstruct the Attacker's movement, based on its belief about the Attacker's abstraction, while the Attacker chooses a trajectory that may intentionally obfuscate its own abstraction of the environment to mislead the Defender. We show that purposeful abstraction concealment naturally emerges from this formulation as a means of improving the Attacker's performance. To solve the game, we propose a bilinear programming approach that integrates Bayesian inference, strategic planning, and belief manipulation. Simulations demonstrate that, by shaping the Defender's belief, the Attacker can induce suboptimal Defender barrier placement, thereby gaining a strategic advantage.