Ansicht
Dokumentation
Aggregated Long-Term Forecast ( RELNBW_702_BCT_PLA_RMP )
BAL Application Log Documentation PERFORM Short ReferenceThis documentation is copyright by SAP AG.
Short text
Aggregated Long-Term Forecast
Use
As of SAP Netweaver 2004s BI Content Add-On 2, you can use the aggregated long-term forecast, which is intended for use in the area of retail.
One of the most important processes in retail is materials planning, in which you create planning data for sales, goods receipts, stock, and so on for future periods.
These planning values are normally specified at the following levels:
- At company or sales organization level (with strategic alignment)
- At the level of the relevant stores (with alignment to the sales view)
- At different material levels (with alignment to the purchasing view)
The materials plan is a key figure plan, that is to say, planning data is entered for combinations of characteristics (for example, store, material group, and article hierarchy level) and compared at different levels.
To minimize the processing effort for planning data in Retail planning, use the aggregated long-term forecast, which you can use to create proposed planning values for the key figures to be planned on the basis of the historical actual data in the selected planning horizon.
Features of the aggregated long-term forecast:
- The forecast is regression-supported.
- The forecast takes the external influences into account using "influencing factors".
- New dialog for processing freely definable events (special types of influencing factors) and their temporal occurrences.
- The forecast level can be selected flexibly.
- The forecast supports the periodicities week, month and posting period.
- The forecast is possible for objects without history by using the reference information.
You can use the forecast in the BPS planning framework in the SAP BW.
Procedure
The aggregated long-term forecast is based on a regression-supported mathematical model.
To be able to execute a forecast on the basis of this model, you must execute model training.
In
model training, weights (values of regression coefficients) are calculated from the historical data
for the key figure to be forecast and the historical values of the influencing factors using an analysis process with a regression analysis, and then transferred to special data containers for long-term storage.
These regression coefficients are used for the forecast in the following three steps:
-
Cleaning up the known influences in the historical growth of the key figure to be forecast in the relevant period using the results of model training.
Result:
- The remaining element of the key figure that cannot be explained by influencing factors.
-
Calculation of the future element of the key figure that is not triggered by influencing factors, by
using the time-series-based methods of a statistical forecast and, if necessary, taking the trend, season, and trend-season models into account.
Result:
- The forecast element that cannot be explained by influencing factors of the key figure in the future.
-
Enrichment of the calculated element of the key figure by adding the effects of the influences that will occur in the future using the result of model training.
Result:
- The future course to be forecast of the key figure to be planned (plan default data).
You can execute steps 1 to 3 as often as required or as often as the plan revision. You can carry out model training far less frequently than the actual forecast.
As a result, you receive plan default data for a key figure for which you want to determine plan values in merchandise and assortment planning.
Effects on Existing Data
Effects on Data Transfer
Effects on System Administration
Tasks for the system administrator:
- You must ensure the continuous update of the actual data for the key figure(s) to be forecast.
- You must ensure the provision of values for the influencing factors that are to be taken into account.
You can ensure the cyclical supply of planned default data to merchandise and assortment planning by running the aggregated long-term forecast in background processing.
Effects on Customizing
Further Information
ABAP Short Reference rdisp/max_wprun_time - Maximum work process run time
This documentation is copyright by SAP AG.
Length: 5017 Date: 20240603 Time: 030724 sap01-206 ( 82 ms )