📝 SummarXiv

Abstract

In this paper, we present a novel, game-theoretic model of deception in two-player, zero-sum games. Our framework leverages an information asymmetry: one player (the deceiver) has access to accurate payoff information, while the other (the victim) observes a modified version of these payoffs due to the deception strategy employed. The deceiver's objective is to choose a deception-action pair that optimally exploits the victim's best response to the altered payoffs, subject to a constraint on the deception's magnitude. We characterize the optimal deceptive strategy as the solution to a bi-level optimization problem, and we provide both an exact solution and an efficient method for computing a high-quality feasible point. Finally, we demonstrate the effectiveness of our approach on numerical examples inspired by honeypot deception.