A fuzzy solution approach for the roster planning problem

The airline crew rostering problem is considered from a dynamic point of view. The study tackles disturbances which lead to modifications for the monthly personalized assignment of the crew staff to flights. An adaptive solution approach is proposed through the asynchronous resolution of an updated optimization problem whose objective is to minimize changes from the previous crew staff assignment while covering the flights of a sliding time period. A mathematical formulation, including penalties which are applied to the assignment changes, is developed. Since the discomfort caused by these changes is often difficult to quantify, a fuzzy logic approach is developed to realize this evaluation. Then, a numerical solution approach based on a forward dynamic programming technique is described. The proposed solution strategy is applied to a medium size problem