Team formation with learning agents that improve coordination

Learning agents increase their team's performance by learning to coordinate better with their teammates, and we are interested in forming teams that contain such learning agents. In particular, we consider finite training instances for learning agents to improve their coordination before the final team is formed. We formally define the learning agents team formation problem, and focus on learning agent pairs that improve their coordination. Learning agent pairs have heterogeneous rates of improving coordination, and hence the allocation of training instances has a large impact on the performance of the final team.