A variation of the network flow algorithm to optimize the diversity of project groups

Abstract

Research has shown that the amount of diversity within project groups has an effect on how well groups perform. In particular, groups that are balanced with respect to gender, race, national origin, and social background tend to be more creative, more productive, and perform better on average than non-diversified groups (Hunt 2015, Hunt 2018). Therefore the assignment of employees to project groups must also take into account the sub-goal of diversification during the decision making process (in addition to hard constraints such as suitable skill level of an employee to even qualify for membership in a group.) The additional constraints this imposes on the optimization problem of finding the best group composition can be a considerable challenge for management. In this paper, we present a new variation of a classical network flow algorithm to solve this problem optimally. We show the results of an application area where optimal groups have been formed that work together. The results have been obtained by the software implementation of our algorithm called "NF Group Diversity".