Research on Multi-Target Allocation Algorithm for UAV Swarms Based on IA2C Reinforcement Learning

In the target distribution area, unmanned aerial vehicle (UAV) swarms perform tasks of discovering, identifying, and capturing mobile target groups. The key link is the effective allocation of target groups. This study proposes an automatic target allocation strategy for UAV swarms based on the independent advantage actor-critic (IA2C) reinforcement learning algorithm, aiming to enhance the autonomous allocation ability of UAV swarms. IA2C is scalable in a distributed environment and effectively solves the problems of multi-agent collaboration and competition by enabling independent learning for each agent. Through this method, UAV swarms can efficiently strike more targets within a limited time while reducing the risk of being intercepted, thereby maximizing the overall strike efficiency.

I. Introduction

In the modern battlefield environment, heterogeneous unmanned aerial vehicle (UAV) swarms undertake the complex tasks of rapidly discovering, accurately identifying, and effectively attacking groups of moving targets. These tasks not only require the UAV swarms to have a high degree of autonomy but also to cope with countermeasures such as “interception, communication jamming, and deployment of decoy targets” that the targets may adopt. To efficiently strike more targets within a specified time and minimize the risk of interception, this study proposes an automatic target assignment strategy based on the Independent Advantage Actor-Critic (IA2C) algorithm, aiming to significantly enhance the combat efficiency and survivability of UAV swarms [1–2].