Distributed maze exploration using multiple agents and optimal goal assignment