Research Of Decision-making Of Platoon Agent In Robocuprescue Simulation System

RoboCupRescue Simulation System is a multi-agent system which simulates the real situation of urban earthquake, Agents search and rescue. It is characterized by heterogeneity Agents, complexity task relation, dynamic environment, tasks allocation and collaboration. The rescue strategy of Agents in RoboCupRescue Simulation System will directly contribute search and rescue problems in actual disaster, which has very far-reaching and of great significance.In this dissertation, on the basis of collation and analysis of relevant literature data, research Decision-making of Platoon Agents on the platform of RoboCupRescue Simulation. The primary research is as follows:1. Decision-making of Police Force.Police Force cleans roads. Implement Decision-making of Police Force based on Utility Function, and implement Platoon Agent to path search based on A*algorithm. It is verified the effectiveness of the strategy by experiment.2. Decision-making of Ambulance Team.Ambulance Team rescues the wounded Civilian and Platoon Agent. Use Particle Filter algorithm to predict lifetime of Civilian, sort rescue order according to Civilian’s lifetime, MUMI implement Decision-making of Ambulance Team. It is verified the effectiveness of the strategy by experiment.3. Decision-making of Fire Brigade.Fire Brigade extinguishes the fire. Cluster buildings through the K-means algorithm, Cluster fires through the incremental K-means algorithm, implement Decision-making of Fire Brigade based on partition. It is verified the effectiveness of the strategy by experiment.At the end of this dissertation, the main research is summarized. Problems which need to further research.