Oracle Retail Knowledge Base

Set up the LPO forecast

2 min read

Context

LPO optimizes against forecast demand parameters, so the forecast has to be trained, approved and mapped to LPO before any optimization run produces recommendations. Done in the AIF UI, as part of Build the LPO application.

Navigate Strategy & Policy Management > Manage Forecast Configurations.

Check the configuration flags before aggregating

RSE_INV_WHSE_ACTIVITY_USE_FLG must be N before starting aggregation — leaving it Y increases aggregation time dramatically, and it can be turned on later once batches are running. Also confirm PMO_PROD_HIER_TYPE (3 for extended hierarchy) and PMO_AGGR_INVENTORY_DATA_FLG in RSE_CONFIG, reachable from AI FOUNDATION CLOUD SERVICE > TOP LEFT MENU > Tasks > Control and Tactical Center > Manage Configurations.

Set up the forecast

  1. Create a forecast run type at the recommendation level you want, for example Style-Color, Price Zone, Week.
  2. Click Start Data Aggregation. Populates PMO_ACTIVITIES and PMO_CUM_SLS.
  3. Create the forecast run under that run type and Submit to train the model.
  4. Review the resulting demand parameters.
  5. Approve in order: Approve Demand Parameters, Approve Base Demand, Approve Forecast.
  6. Activate the forecast run type and map it to LPO.

Validate

select * from PMO_ACTIVITIES;
select * from PMO_CUM_SLS;

Both empty after step 2 means the aggregation found no data — check that the historical facts landed, in Validate AIF Apps historical facts.

Reference

Graph View

Backlinks