The most commonly used replenishment models are all variants of the basic min-max models and assume the demand is normally distributed, meaning there is some level of demand regularly. Examples include Fixed Order Cycle Planning, Fixed Order Quantity Planning and Distribution Requirements Planning.

Recently however demand has become less predictable as a result of all the disruptions. We see more and more examples of lumpy demand, which requires a different approach.

Through various simulation exercises, my conclusion is that the FILL-UP MODEL should be added to the mix and applied in those cases where both the timing and the quantity of the demand is irregular. Using a fill-up model leads to a much better trade-off between inventory investment and service, is easy to set up and does not rely as much as the min-max models on a reliable, unbiased forecast in monthly, weekly or daily buckets.