Architecture of an AI-Based Automated Course of Action Generation System for Military Operations

In a new paper published on arXiv (2604.20862), researchers Ji-il Park, Inwook Shim, and Chong Hui Kim propose a novel architecture for an AI-based automated Course of Action (CoA) planning system tailored for military operations. The study highlights that as warfare evolves—with faster maneuver speeds, longer surveillance ranges, and expanded weapon reach—the operational area grows, making traditional manned CoA planning increasingly impractical. To address this, the authors aim to automate the planning process using AI, integrating relevant doctrines and advanced AI technologies at each stage of CoA generation. The paper includes 15 figures and 2 tables, providing a detailed framework for development.

The proposed architecture is designed to handle the complexity of modern battlefields, where rapid decision-making is critical. By leveraging AI techniques such as multi-agent systems (as indicated by the cs.MA subject classification), the system can generate and evaluate multiple courses of action efficiently. The researchers note that while several countries and defense organizations are actively developing similar systems, security restrictions and limited public disclosure make it difficult to assess technical maturity. This paper aims to bridge that gap by presenting a publicly accessible blueprint that could accelerate the development of automated CoA systems, ultimately enhancing military responsiveness and strategic effectiveness in future conflicts.