A Generic Metaheuristic Approach to Sequential Security Games

TL;DR

This paper presents EASG, a versatile evolutionary algorithm for solving Sequential Security Games efficiently and reliably, capable of handling large instances and providing good solutions quickly in time-critical scenarios.

Contribution

The paper introduces EASG, a general, game-independent evolutionary algorithm for Sequential Security Games that outperforms existing methods in scalability and efficiency.

Findings

EASG achieves near-optimal solutions across diverse game types.

EASG scales better than state-of-the-art approaches.

EASG's anytime feature makes it suitable for time-critical applications.

Abstract

The paper introduces a generic approach to solving Sequential Security Games (SGs) which utilizes Evolutionary Algorithms. Formulation of the method (named EASG) is general and largely game-independent, which allows for its application to a wide range of SGs with just little adjustments addressing game specificity. Comprehensive experiments performed on 3 different types of games (with 300 instances in total) demonstrate robustness and stability of EASG, manifested by repeatable achieving optimal or near-optimal solutions in the vast majority of the cases. The main advantage of EASG is time efficiency. The method scales visibly better than state-of-the-art approaches and consequently can be applied to SG instances which are beyond capabilities of the existing methods. Furthermore, due to anytime characteristics, EASG is very well suited for time-critical applications, as the method can be terminated at any moment and still provide a reasonably good solution - the best one found so far.