Multiagent Coordination by Auctioning POMDP Tasks
Matthijs T. J. Spaan, Nelson Gonçalves, João Silva Sequeira · 2009
We present a procedure for the estimation of bids in auction protocols. This class of protocols is of interest to multia-gent systems because they can be used to coordinate the assignment of tasks to agents. The main idea is to take ad-vantage of methods for the synthesis of task execution con-trollers that rely on the availability of value functions. These provide a natural way to obtain the bid values for a given task. The approach is demonstrated on an active surveil-lance system, where mobile robots must approach and iden-tify humans, and conduct patrols. The Partially Observable Markov Decision Process (POMDP) framework is used to compute policies for the execution of tasks by each agent, the task bid values are obtained directly from the respective value functions. Several simulation examples are presented for an urban surveillance environment, illustrating the ap-plicability of our ideas.