Abstract:

To address the issues of plan rigidity, response delay, and inefficient coordination in the traditional prediction-response paradigm in dealing with complex dynamic disaster chains, this study proposed a new intelligent emergency paradigm capable of real-time perception, dynamic deduction, and adaptive decision generation. A scenario-response decision-making model based on an MAS was constructed. A collaborative framework composed of a commander agent, an information perception agent, a disaster simulation agent, a decision support agent, and a human-machine collaboration agent was established. An agent collaboration mechanism driven by an event-task chain with a digital twin as the shared space and human-in-the-loop as the core was designed. The model was verified using a coal mine gas explosion chain disaster as an example. The results show that the model can form a complete dynamic closed loop of perception-reasoning-decision-making-coordination, which significantly improves emergency response speed, resource allocation efficiency, and personnel safety protection, thereby achieving a paradigm shift from static plan matching to dynamic scenario-based response. The model can significantly improve emergency response speed, resource utilization efficiency, and personnel safety assurance levels.

Key words: emergency decision-making, multi-agent system (MAS), scenario-response, disaster chain, digital twin, collaborative mechanism