Optimally Auditing Adversarial Agents

2026-04-27

AI Summary

This study addresses the problem of strategic misrepresentation by agents in resource allocation settings—such as welfare or credit fraud—by formally modeling the design of optimal auditing strategies in a multi-agent adversarial environment for the first time. The problem is cast as a Stackelberg game, wherein a principal commits to an auditing policy and multiple self-interested agents respond by selecting equilibrium behaviors that are most detrimental to the principal’s objectives. The work distinguishes between adaptive and non-adaptive auditing mechanisms and incorporates realistic audit budget constraints. Leveraging tools from game theory, mechanism design, and equilibrium analysis, the authors develop efficient algorithms to compute optimal auditing strategies under both settings and extend these methods to scenarios with limited auditing budgets, achieving a principled balance between theoretical rigor and practical applicability.

Abstract

Fraud can pose a challenge in many resource allocation domains, including social service delivery and credit provision. For example, agents may misreport private information in order to gain benefits or access to credit. To mitigate this, a principal can design strategic audits to verify claims and penalize misreporting. In this paper, we introduce a general model of audit policy design as a principal-agent game with multiple agents, where the principal commits to an audit policy, and agents collectively choose an equilibrium that minimizes the principal's utility. We examine both adaptive and non-adaptive settings, depending on whether the principal's policy can be responsive to the distribution of agent reports. Our work provides efficient algorithms for computing optimal audit policies in both settings and extends these results to a setting with limited audit budgets.

Problem

• fraud
• adversarial agents
• audit policy
• principal-agent game
• resource allocation

Innovation

• audit policy design
• principal-agent game
• adversarial agents
• adaptive auditing
• limited audit budget