|
Abstract:
|
This paper presents a team of agents for the RoboCup Rescue Simulation League problem that uses an evolutionary reinforcement learning mechanism called XCS, a version of Holland's Genetic Classifiers Systems, to support the agents' decision process. In particular, we use this mechanism to decide the number of ambulances required to rescue a buried civilian and the number of Fire Brigades necessary to extinguish a fire. We also analyze the problems implied by the rescue simulation and briefly describe our solutions for every identified sub-problem using multi-agent cooperation and coordination built over a subsumption architecture. Our classifier systems were trained in different disaster situations. Trained agents outperformed untrained agents and most participants of the 2004 RoboCup Rescue Simulation League competition. This system managed to extract general rules that could be applied to new disaster situations.
|
