Below is a simple graphical representation of a spare parts supply chain.
Currently there is a centralized spare part inventory where ROP is calculated based on outgoing demand. SS= normsinv(0.99)*Stdeviation* Sqrt(Lead time demand + review period).
Since this is only assuming the variability in outgoing demand, it does not actually consider the variability in demand from the service centres. In the future I would like to decrease the safety stock required by considering local DC variability in demand, but there is a problem I need to solve first:
The spare part demand for majority of spare parts is not normally distributed, so the z-score currently used does not seem to be reliable. This is further backed up by the fact that the spare part availability is between 85% to 87% at the service centres.
Are there any substitutes or ways to calculate safety stock when the demand is not normally distributed? for instance, alternative to z-score?
best regards,
Janis