We study the problem of selecting an optimal set of agents in two special cases. In case 1, agents produce individually but also contribute to the output other agents. That is, each agent is described by three characteristics: individual productivity, the ability of helping others, and the cost of hiring the agent. In case 2, the hired agents divide themselves into smaller teams in which they produce output, like researchers divide themselves into coauthoring groups. We develop a model of the way in which agents are divided into smaller teams, mirroring joint research.
We provide algorithms for selecting optimal sets. In particular, (1) the optimal sets may not include agents whose output exceeds the cost of hiring them; (2) the composition of the optimal team is more sensitive to the changes in the cost of hiring agents who mainly contribute to the output of others than the cost of hiring agents who mainly contribute individually.