An Efficient Multi-Agent Collaboration Algorithm for Battlefield Kill Chains via Deep Reinforcement Learning

In modern warfare, rapid and precise battlefield decision-making and effective resource allocation are critical for enhancing operational efficiency, making the construction of an efficient kill chain a key challenge. This paper proposes a deep reinforcement learning-based multi-agent collaboration algorithm for kill chain construction tasks. The battlefield environment is modeled as a multi-agent system, encompassing reconnaissance, decision-making, attack, and logistics. Within the system, agents are assigned specific roles and are trained based on the Actor-Critic framework. Through a collaborative learning mechanism, agents can dynamically adjust strategies to optimize global objectives. Experimental results demonstrate that the algorithm significantly improves task completion efficiency, resource allocation, and decision accuracy, offering a novel solution for decision support in complex battlefield environments.