An Initial Study to Use Deep Reinforcement Algorithm to Improve Efficiency of Emergency Response

Emergency decision making problem is an significant problem in emergency response area. Existing solutions for emergency decision making usually seek for the most similar historical emergency response events with current one, then apply the corresponding response process, which lacks of real-time capability. With the development of machine learning, reinforcement learning stands out for its strong learning ability and wild application in decision making. This paper proposed an approach that translates emergency decision making problem into reinforcement learning problem, and developed an agent can generate optimized strategy for emergency decision making using deep reinforcement learning method.