Fast and the Furious: Hot Starts in Pursuit-Evasion Games

2025-10-12

AI Summary

In pursuit-evasion games, the lack of prior positional information about evaders hinders efficient initialization of pursuer deployment. Method: This paper proposes an end-to-end strategy initialization framework integrating game-theoretic modeling with graph neural networks. Its core innovation is a Pareto-optimal configuration framework in graph feature space, where a Graph Convolutional Network (GCN) learns a “warm-start” pursuer layout optimized for containment efficacy. Multi-objective optimization—minimizing pursuit distance, maximizing containment coverage, and accelerating evader survival-rate decay—is embedded directly into the graph representation learning process to enable rapid response under dynamic adversarial conditions. Results: Experiments demonstrate that, compared to random initialization, the method significantly reduces average evader survival time (−38.2%), decreases total pursuer travel distance (−29.5%), and improves containment success rate (+22.7%).

Technology Category

• Search and Optimization: Learning to Search
• Game Theory and Economic Paradigms: Adversarial Learning
• Constraint Satisfaction and Optimization: Constraint Learning and Acquisition

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems

Abstract

Effectively positioning pursuers in pursuit-evasion games without prior knowledge of evader locations remains a significant challenge. A novel approach that combines game-theoretic control theory with Graph Neural Networks is introduced in this work. By conceptualizing pursuer configurations as strategic arrangements and representing them as graphs, a Graph Characteristic Space is constructed via multi-objective optimization to identify Pareto-optimal configurations. A Graph Convolutional Network (GCN) is trained on these Pareto-optimal graphs to generate strategically effective initial configurations, termed "hot starts". Empirical evaluations demonstrate that the GCN-generated hot starts provide a significant advantage over random configurations. In scenarios considering multiple pursuers and evaders, this method hastens the decline in evader survival rates, reduces pursuer travel distances, and enhances containment, showcasing clear strategic benefits.

Problem

• Optimizing initial pursuer positions without evader location knowledge
• Developing game-theoretic GCN approach for strategic configuration generation
• Enhancing pursuit efficiency through Pareto-optimal hot start configurations

Innovation

• Combines game theory with Graph Neural Networks
• Constructs Graph Characteristic Space via multi-objective optimization
• Uses GCN to generate strategic hot start configurations