A multi-objective formulation of the team formation problem in social networks
Julio Juárez, Carlos A. Brizuela · Proceedings of the Genetic and Evolutionary Computation Conference · 2018
The Team Formation Problem in Social Networks (TFP-SN) consists of finding a team of experts, from a social network, that better undertake a given task. It is mandatory for the team to meet the skill set required by the task and it is desired that the team members communicate effectively to achieve their goal. This problem was proven to be NP-hard for the optimization of different variants of a communication cost function. Even though, real-life instances of this problem involve the simultaneous optimization of two or more conflicting objectives, the studies of the TFP-SN under the multi-objective model has been rather scarce. In this work, we introduce the TFP-SN as a multi-objective optimization problem for the maximization of two conflicting objectives, the collaborative density and the team's ratio of expertise. We tackle this problem employing the NSGA-II framework, for which a proper representation and variation operators are proposed. Experimental results show that the proposed approach generates competitive solutions when compared with well-known heuristics for this problem. Additionally, as a response to the lack of benchmarks and to setup a baseline for future comparisons, we provide a detailed description of the generated instances.