Hierarchical Decision Making for Search and Rescue Teamwork

This paper focuses on the development of a novel approach for modeling the behavior of agents in the RoboCup Rescue Agent Simulation Competition. A two layered approach is developed, with a macro-level behavior responsible for the strategic, high level decisions and a micro-level behavior dealing with the local particularities of the competition. The micro-level behavior was implemented using a simple Markov Decision Process. For the macro-level behavior a Bayesian Game Approximation was compared against a Distributed Constraint Optimization problem formulation. Experimental results show a good overall performance of the methods considered, with their differences being better highlighted on harder, custom made configurations of the official contest maps.