Forecast Engine: Enhancement to aggregation of Do Not Fit flag

The Forecast Engine is enhanced in relation to the ‘Do Not Fit’ input measure to support aggregation where the forecast engine is executed at an aggregate item or aggregate location level.

Previously, while calculating at a level of aggregation, where Item Level or Location Level are not the base level; the mapped measure Do Not Fit has been aggregated from base elements to the level of forecasting. Therefore, aggregates can be excluded from forecasting, but this has been based on any base-level children set to ‘do not fit’.

This process is enhanced such that any base item-location set to Do Not Fit is now excluded from aggregation to the group where forecasting at an aggregate item or location. Any existing output measures from the forecast are not cleared or overwritten by forecast engine results.

For example:

  • If item ‘Car’ at Location "North" has three elements (Item X, Item Y, Item Z) with history values for a period: 100, 100, 100.
  • ‘Item Y’ has Do Not Fit measure with a value >0.
  • Aggregated value of ‘Car’ excludes the value of ‘Item Y’ @ ‘North’ and consider a value of 200, by summing the values of Item X and Item Z.
  • The resultant forecast for ‘Car’ is 120.
  • The forecast persisted to elements of Item ‘Car’ are:
    • Item X forecast value =60
    • Item Y forecast value =0
    • Item Z forecast value =60
  • The same process is applied for the replacement history engine, to output replacement history values.

To exclude an aggregate item-location from being forecast, or updated by the engine, the Do Not Fit flag can be set at the aggregate level and spread to all base combinations, or all base item-location combinations can be set to Do Not Fit > 0.

This does not affect the period level of forecasting, since the flag is not time dependent.

Note: 
  • Do Not Fit logic is not supported when calling the engine with "Spreading Measure for time-phased results" defined - results are prorated to base cells irrespective of Do Not Fit. This logic can be used with the default proration based on ‘Spreading Measure for period constant results’.
  • Do Not Fit logic is not supported when calling the CallForecastEngine macro with the Filter parameter populated. The forecast engine output measures are cleared for all item-location combinations passed to the engine irrespective of the Do Not Fit value. This value can be used without the Filter parameter defined, or when calling the forecast engine from the worksheet for the current context (Fit Forecast using default engine, Fit Forecast with options).

By default, this feature is enabled. No additional role or privilege is required to access this feature.