Development and evaluation of a team formation algorithm for gaming environments

In the context of multiplayer online games, but in many other group activities that entail collaboration between user, appropriate group formation is a critical issue. Groups should be formulated in a way that on the one hand user collaboration (or collaboration avoidance) preferences are respected, and on the other hand collaboration potential and productivity of groups is maximized.

In the current thesis, we consider an internet-based collaboration environment, where users work in pairs and, after their collaboration, each of them rates the quality of work and the quality of collaboration of his/her partner. Moreover, users have the option to express a positive, negative or neutral opinion towards the possibility of cooperating with individual users. For such an environment, we have developed an algorithm that employs linear
programming techniques, in order to create an optimal user pair formulation, considering the ratings that users have assigned to other users and collaboration intention statements. We present in detail the system structure, the algorithms and techniques employed, and we analyze the performance of the system in terms of the time needed to formulate user pair assignments
in various indicative cases. Finally, potential future extensions are outlined.