Uncertainty Based Hybrid-intelligence Multi-branch Wargaming

Uncertainty of war challenges command decision making, especially in the pre-war preparation stage. In newly proposed operational concepts, uncertainty is taken as a means to limit the opponent's decision-making. How to help commanders analyze and utilize uncertainty has become a technical challenge that must be overcome to win the future combats. In this paper, a method called hybrid-intelligence multi-branch wargaming was proposed. Through combination of human intelligence in form of knowledge, and machine intelligence trained by reinforcement learning, the method realized new functions of COA evaluation and optimization analysis. Its feasibility and effectiveness has been verified through prototype development and experiments. It provides a short-term feasible way for the application of artificial intelligence technology in the field of command decision-making.