Research Of Cooperation Method Of Multi-agent In Robocup Rescue Simulation

In recent years, in order to promote the artificial intelligence technology in the application field of social development, the RoboCup RoboCupRescue project produced closer to applications. As a new research field of RoboCup, RoboCupRescue mainly geared to the needs of the development of the social emergency rescue system, which includes two aspects of hardware and software. RoboCupRescue’s main goal is to build a rational intelligent search and rescue and dispatching system, and the disaster comes, make social loss reduce to a minimum as far as possible (make the highest score in the game).In the RoboCup rescue simulation system, multi-agent cooperation on overall rescue’s effect plays a vital role. In the collection, sorting and analyzing related literature, on the basis of taking the RoboCup rescue simulation system for research and experimental platform, multi-agent cooperation method was studied, the main research content is as follows:(1) The cooperation method of police force Agents.In the RoboCup rescue simulation system, the task of cooperation between the police Agent is good or bad directly affects the overall efficiency. In without police center how to carry out tasks, a task priority cooperation method is put forward. Through to compare and choose the task priority, improve police Agent for the completion of important tasks of cooperation for more efficiency. Finally through the simulation experiments verify the effectiveness of the proposed cooperation method in rescue simulation system.(2) The cooperation method of ambulance Agents.Ambulance agents need timely rescue injured people. Int order to rescue in the shortest time, a method is used based on distance calculation to achieve the optimal allocation for rescue missions. Through the simulation experiments verify the effectiveness of the proposed cooperation method in rescue simulation system.(3) The cooperation method of fire brigade Agents.In the case of have fire center and more fire disaster, need the fire center through the global communication collect local information and fuse global map information, and make a reasonable fire task distribution points. Aiming at how to form efficient cooperation in global map, put forward a kind of Agent based utility function value and genetic computation task allocation method. This method can effectively solve the optimal task allocation scheme, finally improve the game scores.At the end of this dissertation, the main research is summarized. It makes out the points of innovation and main research results, also give the shortage in the dissertation and the problem need to further research.