A column generation approach for the shift design and rostering problem in airport ground handling

In this study we propose a Column Generation based solution method that simultaneously solves a shift design- and rostering problem. In our problem setting, we consider a heterogeneous workforce demand of ground handling staff that is derived from a flight schedule in a given planning horizon. The pricing problem amounts to finding a shift schedule of a worker that corresponds to a longest path problem in the corresponding graph. In our case, several labour rules should be respected in the worker schedules which brings extra complexity to the pricing problem. Besides worker availabilities, we also consider worker preferences while constructing schedules. The instances contain around 200 workers and span a one-month planning horizon. Two types of workers are considered: First, we have workers with a fixed target amount of working hours each month. Second, we consider flexible workers that have no required target of working hours, but have a preferred number of working hours that can change each month based on preference. Each worker, when on duty, can cover the demand of various workloads. Our solution approach produces timetables and employee schedules with the goal of minimizing all labour costs while satisfying the workload demand. We introduce heuristics that employ graph theory for the pricing problem. Our approach is successfully tested on real life data of a ground handling company in an airport situated in the Netherlands.