Memory Assignment for Finite-Memory Strategies in Adversarial Patrolling Games

2025-05-20

AI Summary

In adversarial patrol games, allocating memory resources among states for finite-memory strategies—aiming to minimize the attacker’s worst-case damage—is a challenging optimization problem. Method: This paper proposes the first iterative, general-purpose memory allocation framework that jointly optimizes memory distribution and patrol strategy without manual tuning. It integrates finite-state machine modeling, iterative improvement algorithms, and an end-to-end black-box optimization architecture, supporting diverse patrol models and compatible with arbitrary black-box strategy optimizers. Contribution/Results: Evaluated on multiple benchmark instances, our framework significantly outperforms hand-crafted memory allocations in defense performance. It enables practical deployment of finite-memory strategies by overcoming a longstanding bottleneck in their real-world applicability, establishing a new standard for automated resource-aware strategy synthesis in security games.

Technology Category

• Search and Optimization: Adversarial Search
• Multiagent Systems: Adversarial Agents
• Game Theory and Economic Paradigms: Adversarial Learning

Application Category

• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Security and Privacy: Large-scale security measurements
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Adversarial Patrolling games form a subclass of Security games where a Defender moves between locations, guarding vulnerable targets. The main algorithmic problem is constructing a strategy for the Defender that minimizes the worst damage an Attacker can cause. We focus on the class of finite-memory (also known as regular) Defender's strategies that experimentally outperformed other competing classes. A finite-memory strategy can be seen as a positional strategy on a finite set of states. Each state consists of a pair of a location and a certain integer value--called memory. Existing algorithms improve the transitional probabilities between the states but require that the available memory size itself is assigned at each location manually. Choosing the right memory assignment is a well-known open and hard problem that hinders the usability of finite-memory strategies. We solve this issue by developing a general method that iteratively changes the memory assignment. Our algorithm can be used in connection with emph{any} black-box strategy optimization tool. We evaluate our method on various experiments and show its robustness by solving instances of various patrolling models.

Problem

• Optimizing memory assignment for finite-memory Defender strategies
• Automating memory allocation in adversarial patrolling games
• Improving transitional probabilities between Defender strategy states

Innovation

• Iterative memory assignment optimization method
• Compatible with any black-box strategy tool
• Improves finite-memory Defender strategies robustness