Robust optimal policies for team Markov games