Overview of Forecasting Using Supersession Relationships
You can generate a forecast for the current revision and latest (future) revision of an item in a supersession chain or across multiple supersession chains by using the demand history of previous item revisions.
When the functionality for forecasting using supersession relationships is used, the demand history for previous item revisions and the available demand history for the current item revision from the history start date to the history end date of a plan are aggregated. Then, a forecast is generated on the basis of this aggregated demand history and applied to the current item revision and latest item revision from the plan start date to the forecast end date.
In industries such as automobile manufacturing, the hundreds or thousands of parts that make up a finished end item go through numerous revisions on account of design or production changes. Without the functionality for forecasting using supersession relationships, you would have to manually copy the demand history of previous item revisions to the current revision and latest revision to generate a forecast. Thus, this functionality enables you to improve efficiency by automating the process for forecasting for the current revision and latest revision in a supersession chain.
Points to Note About Forecasting Using Supersession Relationships
Note these points about forecasting using supersession relationships:
- The functionality covers only supersession chains with reciprocal relationships
among the item revisions.
Supersession chains with unidirectional relationships aren't considered.
For example, consider a supersession chain with item revisions A, B, C, and D, in which A and B have a reciprocal relationship, B is superseded by C, and C and D have a reciprocal relationship. Two supersession chains (A < > B and C < > D) are considered. If B is the current revision in the A < > B supersession chain, then B receives a forecast on the basis of the demand history for A and the available demand history for B. Similarly, if D is the current revision in the C < > D supersession chain, then D receives a forecast on the basis of the demand history for C and the available demand history for D.
- Item revisions with start dates that are beyond the plan horizon or forecast end
date won't be considered for forecasting.
For example, if item revision N supersedes item revision M from 01-Sep-2024, and the forecast end date is 31-Aug-2024, then a forecast won't be generated for N.
The start date for an item revision is stored in the Item Revision Start Date measure.
- Providing the end date for an item revision is optional. This date is stored in
the Item Revision End Date measure.
If an item revision has been defined without an end date, then the Item Revision End Date measure for that item revision is populated with the date one day before the start date of the next item revision in the supersession chain. For the latest item revision, the measure is populated with the end date of the plan horizon.
For example, consider a supersession chain with item revisions A, B, C, and D. In this supersession chain, B supersedes A, C supersedes B, and D supersedes C. The start dates for B, C, and D are 01-Feb-2021, 01-Nov-2022, and 01-Jan-2024. The end date of the plan horizon is 31-Aug-2024. For this supersession chain, the end dates set for A, B, C, and D are 31-Jan-2021, 31-Oct-2022, 31-Dec-2023, and 31-Aug-2024.
- When you have a one-to-many supersession relationship in which an item revision
is superseded by more than one item revision, the item revision for the
supersession relationship with the highest rank set in Oracle Fusion Cloud
Product Lifecycle Management receives the forecast.
For example, consider a supersession chain with item revision A that's superseded by item revisions B and C. The supersession relationship between A and B has a rank of one, while the supersession relationship between A and C has a rank of two. For this supersession chain, B receives the forecast that's based on the demand history for A, and C doesn't receive a forecast.
- When you have a many-to-one supersession relationship in which many item
revisions are superseded by one item revision, the demand histories for the
previous revisions across supersession chains are aggregated for the forecast
generation for the current revision and latest revision.
For example, if item revision X is superseded by item revision Z, and item revision Y is also superseded by Z, then the demand history for X and Y is aggregated for the forecast generation for Z.
- You can define supersession relationships in Oracle Product Lifecycle Management
or using the file-based data import (FBDI) template named Supply Chain Planning
Item Substitute.
Note these points about supersession relationships that are defined in Oracle Product Lifecycle Management:
- The item revisions in a supersession chain can belong to different organizations.
- You can collect the item relationships using the Collect Planning Data page. You must select the Item Relationships reference entity on the Reference Data subtab on the Parameters tab.
Note these points about supersession relationships that are defined through the import template
- The item revisions in a supersession chain must belong to the same organization.
- You can collect the item relationships using the Load Planning Data from Flat Files scheduled process.
- To prevent adjusting of the editable forecast measures for the current revision and latest revision outside their effective dates, you can configure an expression using the Item Revision Start Date and Item Revision End Date measures on the Edit Lock tab in the Edit Measure dialog box.
- A forecasting profile that includes the SupersessionForecastLevelMeasure
forecasting parameter can generate a forecast for items with and without a
supersession relationship.
However, for better control over the forecasting process, you may choose to set up one forecasting profile for supersession forecasting and add it to your plan with an analysis set that contains only the items with supersession relationships. You can then set up another forecasting profile without the SupersessionForecastLevelMeasure forecasting parameter for items without supersession relationships and add the forecasting profile to your plan with an analysis set that contains only these items.
- If you use separate organizations for your items with supersession relationships, you can define a plan for these organizations. The organizations containing items without supersession relationships can be covered by another plan.
- During supersession forecasting in demand plans, when the
SupersessionForecastLevelMeasure forecasting parameter is in the forecasting
profiles, relationships among the Item, Organization, Customer Site, and Demand
Class members are created for the current revision and latest revision on the
basis of the relationships among these members for the previous revisions. These
combinations of members for the current revision and latest revision receive the
forecast. The relationships are created for current revisions and latest
revisions that have item revision start dates within 365 days before the plan
start date and item revision end dates after the plan start date. These
relationships are stored in the granularity tables that correspond to the input
measure of the forecasting profile and the forecast status measure.
For example, consider a supersession chain with items A, B, C, and D, in which A and B are the previous revisions, C is the current revision, and D is the latest revision. Item revision A is related to organization M1, customer site S1, and demand class X, while item revision B is related to organization M2, customer site S2, and demand class X. During supersession forecasting, relationships are created for current revision C and latest revision D with organizations M1 and M2, customer sites S1 and S2, and demand class X. These combinations of members for current revision C and latest revision D receive the forecast.
During supersession forecasting in replenishment plans, these relationships are created among the Item and Organization members because the Customer and Demand Class dimensions aren't supported for such plans. The Organization members (locations) can also be subinventories.
- For correct functioning of this feature, the previous revisions, current revision, and latest revision for an item must be present in the plan's scope.
Example of Forecasting Using Supersession Relationships
Consider a supersession chain that consists of item revisions A, B, C, and D. There are reciprocal supersession relationships between A and B starting from February 2021, between B and C starting from November 2022, and between C and D starting from January 2024.
If the plan start date is between November 2022 and January 2024, A and B are the previous revisions, C is the current revision, and D is the latest revision.
When the functionality for forecasting using supersession relationships is used, the demand history for A and B and the available demand history for C from the history start date to the history end date are aggregated. Then, a forecast is generated and disaggregated for C from the plan start date to the end of December 2023 and for D from January 2024 to the forecast end date.
The following figure depicts this example:
