Forecasting spare parts demand for aperiodic inspection intervals

Advanced technical systems are complex machines that consist of several components. For these systems, it is important to forecasts defects and failure, which generate the demand for spare parts. Recently, Wang and Syntetos (2011) introduced a maintenance-based model to forecast the intermittent and lumpy demand for spare parts. In this paper, we tested how well this model performs with aperiodic inspection intervals. We introduce a simple policy to create those aperiodic intervals. It turns out that the maintenance-based model with aperiodic intervals performs as well as with periodic intervals. In most of the cases it outperforms the time series-based benchmark.