Research On Dynamic Assignment And Coordination Of Emergency Task

The emergency has sudden, dangerous and dynamic characteristics. Multi-Agentsystem theory will be used in emergency decision-making, the paper proposes thedynamic task allocation in task allocation layer, dynamic collaboration model in taskexecution according to dynamic environment..Model enriches collaborative approachbased on reinforcement learning method and task allocation approach in multi-Agenttheory and provide an effective solution to the emergency decision-making.The model is divided into two layers: task allocation layer and task execution layer.In task allocation layer, there is the game in system due to constraints and co-contradictoryrelationship between the Agents, so the article proposed dynamic task allocation based onthe game; in task execution layer, on the one hand, a single Agent may can not completethe complicated task without help of Agent. On the other hand, the multi-Agent dynamicmay be conflict due to incomplete information, of environment Therefore, during in theexecution layer of multi-Agent system the Agent need coordination and cooperation withthe other Agents to improve the efficiency of systems. Due to uncertainty, and dynamicenvironment in decision-making process, the Agents in the system need for real-timelearning, The article proposes reinforcement learning methods to conduct an overall studyin model to find the optimal strategy. The paper proposes algorithm of this model andanalysis algorithm simulation results.Based on the characteristics of the emergency decision-making, combination ofdynamic task allocation and action collaboration is an effective solution for emergencydecision-making problems.