Team Formation Dynamics

Abstract

Using data from online courses, we study the dynamics of team formation in online environments. In particular, we observe that the teams formed by online students for completing course projects are homogeneous in terms of age, location and education level but diverse in terms of primary skill. Motivated by the data, we propose a coalitional game that captures the teaming preferences of individuals and show that the core of the resulting game is always non-empty. Even though our proof is constructive, it does not always yield a polynomial-time algorithm. We show that it is NP-hard to find a solution in the core in the general case and propose polynomial-time algorithms for natural special cases motivated by observations of online course data.