UAV Swarm Tactics: An Agent-Based Simulation Study on Resource Allocation and Retreat Strategies in Asymmetric Engagements

Abstract

Unmanned Aerial Vehicle (UAV) swarms are increasingly recognized for their potential in complex operational environments. Effective command and control, particularly concerning resource allocation and tactical decision-making like retreat maneuvers, remain critical research areas. This paper presents an agent-based simulation framework to investigate the interplay between team composition, resource limits, and behavioral thresholds in asymmetric UAV swarm engagements. The simulation features two opposing teams: a blue team employing Ant Colony Optimization (ACO) for reconnaissance and Brown-Robinson game theory for specialized hunter drone tasking, and a red team utilizing Artificial Neural Networks (ANNs) for adaptive roster selection and spawn positioning, coupled with a rule-based tactical decision module. We specifically analyze the impact of three key parameters on the red team's performance: (1) the maximum allowable team weight, (2) a health-based retreat threshold, and (3) a firepower-imbalance-based retreat threshold. By systematically varying these parameters, this study aims to elucidate their influence on mission outcomes, drone survivability, and overall swarm effectiveness, offering insights into robust strategy development for autonomous UAV operations.