Dynamic Assignment Strategies for Multi-Agent Target Defense Differential Games

Abstract

This paper addresses dynamic assignment strategies in multi-agent Target-Attacker-Defender (TAD) differential games, involving an active target, M attackers, and N defenders. These games are challenging due to their continuous-time dynamics and the need for real-time strategy adaptation in complex, interdependent environments. Traditional static assignment methods often result in suboptimal defense outcomes due to their inability to adapt to evolving game states. To address this, we propose a novel dynamic agent allocation strategy that continuously reassigns defenders to attackers in real time, maximizing the likelihood of successful target defense. Our approach combines Linear Quadratic Differential Games (LQDG) with a recedinghorizon framework, enhancing defender coordination and effectiveness. We demonstrate significant improvements over static methods in dynamic, multi-agent defense scenarios. Furthermore, we analyze cases where M = N, M>N, and M<N, comparing the performance of the proposed algorithm under these conditions through simulation results. © 2025 Elsevier B.V., All rights reserved.