Research on a Multi-Agent Fire-Link Construction Algorithm Based on Thinking Chain in Multi-Source Ground Combat

In multi-source ground combat scenarios, the efficient collaboration of multi-agent systems is crucial for completing tasks such as reconnaissance, decision-making, striking, and logistics support. This paper proposes a thinking chain-based multi-agent fire-link construction algorithm to address the bottlenecks encountered by current multi-agent systems in information fusion and rapid decision-making processes. A four-agent collaborative model is designed based on thinking chain theory, which includes reconnaissance, decision-making, striking, and support agents. Each agent utilizes reinforcement learning to autonomously learn and make decisions to maximize overall operational effectiveness. In the fire-link construction process, the reconnaissance agent is responsible for information collection and enemy situation assessment, the decision-making agent generates combat plans based on reconnaissance data, the striking agent executes attack missions according to the decisions, and the support agent provides logistics and resource allocation. Through continuous collaboration and feedback among the agents, the algorithm optimizes each link in the chain, enabling a more efficient response to combat tasks. Experimental results demonstrate that the proposed algorithm outperforms traditional methods in multi-source information fusion and real-time decision-making, significantly improving combat efficiency and response speed.