The new system requires that the demand patterns are carefully examined and suitable
forecasting techniques are then selected correspondingly. For the nine sample items,
the historical demand patterns of seven items contain only a level and the exponential
smoothing is then applied to forecast the demands of these items. The Holt’s method
is adopted to forecast the demands of the other two items which apparently contain a
trend. The system evaluation indicates that the selected forecasting methods in most
instances perform better than the existing system and result in lower forecast errors.
In managing inventory, two continuous review models are proposed; the reorder point
model and the min-max model. The two models are basically identical except that the
reorder point assumes that gradual customers’ demand while the min-max model
allows for the lumpy customer demand. The parameters for executing these two
inventory models are determined on the basis of Economic Order Quantity (EOQ) and
the safety stock required to protect the system against the variability in demand rate
and replenishment lead time. The simulation results reveal that both models would
likely lead to reduced backorder quantity, improved fill rate, thereby significantly
increasing customer satisfaction. However, with larger safety stock required the two
models would be associated with higher cost than the present situation.