A two-phase heuristic and a lexicographic rule for improving fairness in personnel rostering

A fair allocation of workload to people is crucial for securing job satisfaction. Researchers have introduced numerous objectives and algorithms to represent and improve fairness in personnel rostering problems. These approaches should not ignore the roster quality that is influenced by personnel rostering constraints, such as maximum working times, minimum rest times, etc. The present paper proposes a new fairness objective and an effective two-phase heuristic for optimizing rosters, taking into consideration the established personnel rostering constraints and the fairness. The new fairness objective is based on a lexicographic rule that offers a beneficial trade-off between roster quality and fairness. The new heuristic is tested on real world data and the results show that fair rosters can be obtained without significantly decreasing the roster quality.