Partition based multi-police-intelligent-agent task allocation method in RCRSS (Robo Cup Rescue Simulation System)

Abstract

The invention provides a partition based multi-police-intelligent-agent task allocation method in an RCRSS (Robo Cup Rescue Simulation System). The partition based multi-police-intelligent-agent task allocation method comprises dividing a map into partitions which are not more than the number of policemen, adopting a distance based k-means clustering algorithm, serving the distance as a similarity evaluation index, performing modeling analysis on the rescue simulation environment and obtaining the partitions which are smaller than the number of police intelligent agents and allocating a certain number of policemen in every partition and performing scheduling on the police intelligent agents in a cooperation mode that the static allocation is combined with the dynamic adjustment. According to the partition based multi-police-intelligent-agent task allocation method in the RCRSS, the partition based multi-police-intelligent-agent task allocation method can adapt to multiple complex disaster environments, the task of road block removal is well achieved in the limited time, the cooperation efficiency of the police intelligent agents is improved, and the rescue efficiency of the police intelligent agents is effectively improved.

Description

technical field [0001] The invention relates to a partition-based multi-police agent task assignment method in RCRSS. Background technique [0002] RoboCup rescue simulation is a rescue simulation project in the RoboCup competition. It involves distributed artificial intelligence, intelligent robot technology, multi-agent cooperation and other fields. Events and competitions. As the basic platform of the competition, RCRSS is a typical multi-agent system. It simulates the real-life urban disaster environment by computer, and quickly establishes a rational and intelligent system when houses collapse and fire, fire spreads, roads are blocked, and citizens bury their injuries. Search and rescue and dispatch system to minimize social losses as much as possible. [0003] There are six types of controllable agents in the system, among which the police, fire brigade and ambulance belong to the mobile class of agents, which undertake the tasks of clearing obstacles, fire fighting ...

AI Technical Summary

Problems solved by technology

[0009] The core problem of task assignment in RCRSS is to reasonably map the M rescue agents in the disaster environment to N tasks. pf clearing tasks assigned to N pf Under the current allocation mechanism of a police agent, since there is no designated area for each agent to perform tasks, when the simulation map is large and the number of agents is small, there will be two problems: first, for a single agent, there are no The entire map cannot be completely traversed within a limited task period; secondly, in the simulation experiment, it was also found that multiple police agents performed a task at the same time. a waste of resources

Images