Rolling horizon validation

Rolling horizon validation takes the inventory placement determined by safety stock optimization using historical data and sets time phased inventory policy parameters based on forecast data. It uses the historical demand (typically in the Customer Orders table) and the forecast data populated in the Customer Forecasts table. You define the Rolling Horizon Validation Start Date, which is most often the mid-point of the model horizon. When you run Rolling Horizon Validation, safety stock optimization is run from the beginning of the model horizon to the Rolling Horizon Validation Start Date to determine stocking locations. Forecast data from the Rolling Horizon Validation Start Date to the end of the model is used to set time phased inventory policy parameters based on forecast buckets.

You do not need to run Safety Stock Optimization or Rolling Horizon Modeling prior to running Rolling Horizon Validation.

You can review the results of rolling horizon validation in these output tables:

• Rolling Horizon Validation Details - This table provides time series data of inventory levels. The dates are based on the Forecast Time Bucket.
• Rolling Horizon Validation Service Level - This table displays service levels, fill quantity rates and ready rate values by site-product.

The rolling horizon validation workflow

• Set up your model with historical demand (typically in the Customer Orders table) and forecast data populated in the Customer Forecasts table. You should have the data populated for the full model horizon. For example:
  • Assume a model with one year of historical demand data (from 1/1/2018 through 12/31/2018). This data is populated in weekly buckets in the Customer Orders input table.
  • Populate the Customer Forecast input table with weekly buckets of forecasted demand for the model horizon.
• Set the overall model horizon start and end dates, in this case 1/1/2018 through 12/31/2018.
• Enter the Rolling Horizon Validation Start Date on the Rolling Horizon Modeling tab in Inventory Optimization options. This is typically a date in the middle of the model horizon.
• Run Rolling Horizon Validation. This problem type runs safety stock optimization from the start of the model horizon to the Rolling Horizon Validation Start Date, then performs validation on the remaining part of the model horizon based on the forecast data populated.
• Confirm the results in the Rolling Horizon Validation Details output table.
  • In this table, summary results are populated by Date, based on the Forecast Time Bucket on the Safety Stock Optimization: Advanced tab in Inventory Optimization options.

If you have Tableau, you can also review the Rolling Horizon Validation visualizations in the Inventory Optimization Tableau template. The visualizations are most effective when you use s,S as the inventory policy.