Constrained Team Formation Using Risk Estimation Based on Reputation and Knowledge

Abstract

Given a generalized task that requires a different number of experts for various skills, the team formation problem (TFP) in real-world social networks aims to identify a set of experts that have the requisite skills and can collaborate effectively to accomplish the desired task. This paper considers TFP in realistic settings where the team composition must satisfy certain constraints. Sometimes for a task, only certain suitable experts having high reputation in the team of experts is sufficient to achieve the task. Moreover, not all experts having high reputation/ high expertise are always needed or are available for the task. To evaluate this, we propose a genetic algorithm-based model and introduce risk estimation strategies to determine the suitability of team for a particular task. The experimental results establish that our proposed model is useful for TFP in practical scenarios and discovers more coherent and collectively intelligent teams having low inherent risks.