LAUSR

ABSTRACT In shift scheduling, short-term uncertainties in worker availability frequently cause gaps between strategic calculations based on expected values and the actually available workforce. Strategic decisions on a suitable shift pattern do not only have to cope with stochastic deviations but also, the optimal short-term reactions must be taken into account. Here, we present a framework to evaluate given shift patterns regarding their optimal operational consequences under uncertainty. Based on a reduced worker availability due to unpredictable absences, an optimal assignment of heterogeneously skilled workers to functions is determined to staff different functions as best as possible. This reactive workforce assignment model can additionally be used to support optimal short-term reactions. For detailed evaluation, the workforce availability is modelled by stochastic simulation. The key elements of the framework interact cyclically within a rolling horizon approach to anticipate future availability. The practical relevance of the framework for real-world decision problems is demonstrated in an extensive case-study. For the realistic case of a power plant, we compare two given shift patterns regarding reliability and workload balance. Using our framework, the acceptance of a certain shift pattern increases by both workers and management. Furthermore, skill gaps are revealed that could be bridged by training.