Running an Inventory Optimization Operator

Global (Multi-stage) Inventory Optimization

Objective

After completing this lesson, you will be able to understand the SAP Integrated Business Planning logic for multi-stage inventory optimization.

Definition of Global Multistage Inventory Optimization Operator

Configuring the Planning Horizon Parameter for Inventory Optimization

Standard planning areas have planning horizons defined. You may want to do the following:

  • Use planning horizons different than the standard planning area planning horizons

  • Apply different planning horizons at the planning unit level

To use different planning horizons than the standard, you can define planning horizon parameters for inventory optimization planning operators. The planning horizon parameter value is equal to calendar weeks. The following operators support the planning horizon parameter:

  • Multistage Inventory Opt

  • Calculate Inventory Components

  • Single-Stage Inventory Opt

The following table details the planning horizon parameter:

ParameterDescription
PLANNING_HORIZONPositive integer value that represents calendar weeks.

Note

For detailed information about configuring inventory optimization operators, see Inventory Optimization (IO) Operator in the model configuration guide.

Operator Running Sequence

For the most accurate calculation results, there is a suggested running sequence for the operators. The sequence is the same for both batch and simulation modes, and is as follows:

  1. Run the Manage Forecast Error Calculations Inventory Optimization app.
  2. Run the Global (multistage) inventory optimization operator.
  3. Run the Calculate Target Inventory Components operator or run the Decomposed (single-stage) inventory optimization operator.

The following figure illustrates the run sequence:

Overview available planning operators in inventory optimization

Note

When you run inventory operators, if output key figure data does not display for certain time periods but displays for others, verify that week level time periods are seven days long. If you’re using a week level time profile, weeks must contain seven days. If you need to break weeks that cross months, use technical week time profile.

Key Figure Dependencies

The unified planning area delivered as sample content has built in logic where one key figure requires the existence of other key figures. These key figure dependencies that can cause some operators to run incorrectly or fail to run.

For the Global (multistage) inventory optimization operator, the dependencies are as follows:

If the key figure SAFETYSTOCKDEMANDVAR exists in the planning area, the following key figures must also exist in the planning area:

  • SAFETYSTOCKSUPPLYVAR

  • SAFETYSTOCKSERVICEVAR

  • SAFETYSTOCKLOTSIZE

  • AVERAGESERVICELEVEL

If the key figure CALCTLEADTIME exists in the planning area, the following key figures must also exist in the planning area:

  • CALCTLEADTIMEVARIABILITY

  • CALCTMINLOTSIZE

  • CALCTINCLOTSIZE

If the key figure CALCPLEADTIME exists in the planning area, the following key figures must also exist in the planning area:

  • CALCPLEADTIMEVARIABILITY

  • CALCPMINLOTSIZE

  • CALCPINCLOTSIZE

    Attributes as Key Figures and Time Period Ranges

    We recommend that you set the time period range for attributes as key figures to be from 0 (zero) to at least 13 weeks.

    You will need to reload data at the frequency you use for the range. That is, if you set the range to 13 weeks, you will need to reload data every 13 weeks.

    The best practice is to set the time period range to 13 weeks (or fewer), but if you choose not to update that frequently, we suggest you update every 26 weeks.

    Global (Multistage) Inventory Optimization Operator

    You use the Global (multistage) inventory optimization operator to recommend safety stock across all products and locations of the supply chain. The optimization minimizes your total safety stock holding cost while ensuring that all customer service level targets are met. The operator performs the following:

    • Optimizes safety stock globally and simultaneously across all products and locations of a supply chain while considering demand uncertainties, supply uncertainties, supply quantity, lead times, costs, and service levels.

    • Propagates forecast and forecast variability to stocking nodes that have customer demand.

    • Propagates forecast and forecast variability to internal (upstream) nodes of a supply chain network.

    • Optimizes internal service levels between internal (upstream) nodes of the supply chain network.

    • Calculates safety stock targets, backorders, and average expedited quantities.

    • Calculates values for the whole supply chain network.

    • Supports subnetwork/planning unit (PLUNITID) filters in batch mode.

    • Supports unit of measure conversion for input and output key figures.

    The Global (multistage) inventory optimization operator considers the aggregated demand and outgoing backlog when recommending safety stock in a time-varying binary sourcing environment (that is when the lead time at one supplier is extensively longer than the lead time of the other supplier).

    The operator also considers smoothing for continuous binary sourcing. Spikes and dips in supply are detected based on the sum of the incoming supply ratios offset by their respective lead times. A spike in a period is removed using the maximum arc’s recommended safety stock for that period. A dip is removed using the average of the recommended safety stock before and after that period. Binary sourcing input data must exist in each period of the planning area’s planning horizon configuration for smoothing to work.

    The operator uses the following push logic for non-stocking nodes: for the relationship Stocking Node A → Non-Stocking Node B → Stocking Node C, Lead Time and Lead Time Variability between Stocking Node A and Non-Stocking Node B is pushed to Stocking Node C. The higher the Lead Time or Lead Time Variability, the higher the Recommended Safety Stock at Stocking Node C. The Recommended Safety Stock at Stocking Node C is aligned with pure push logic (that is, cumulative lead time is considered).

    Note

    Stocking Node A can be an internal or external (vendor) node, and supports BOM in non-stocking components. If the non-stocking node is multi-sourced or multiple components in a bill of material, then the lead time is not pushed.

    The Global (multistage) inventory optimization operator rounds the fractional exposure period up to a multiple of weeks in the calculation of the recommended safety stock due to service variability. For example, if periods between replenishment (PBR) = 1 day and lead time = 1 week, the exposure period of 1.14 is rounded up to 2 weeks. The recommended safety stock due to service variability is then calculated based on the demand over the past two weeks.

    The following figure shows the stages of the Global (multistage) inventory optimization operator and what inputs are used:

    Inputs of Global Multistage Optimization Planning Operator
    Outputs of Global Multistage Optimization Planning Operator

    The following table lists the key figure outputs for the Global (multistage) inventory optimization operator:

    The following table lists the Location Product master data type attribute outputs for the Global (multistage) inventory optimization operator:

Note

To run the inventory operators, specific technical IDs defined by SAP must be used for the relevant master data types, attributes, and key figures. If these technical IDs are not used, the inventory operators will fail.

Series or Tandem Supply Chain

The tandem or series supply chain has the following features:

  • Simplest model

  • Node versus stocking point

  • Single versus multi-stage (echelon)

  • Upstream versus downstream

  • Pull and demand propagation

  • Calculation (single-stage)

  • Optimization (multi-stage)

The figure describes the Series Supply Chain.

Process and Distribution Supply Chains

Compare the two supply chain concepts depicted in the following figure to get an impression of what the different supply chains look like.

The figure describes the Supply Chains.

Supply Chain Network Representation

The following figure is the representation of the actual facilities and distribution options. It is a traditional modeling approach.

The figure describes the Supply Chain.

Here is how to represent the actual facilities in an accurate inventory model, using the appropriate symbols.

The figure describes the Inventory Model.

Multi-stage Dilemma

The following figure shows a customer-facing (CF) node satisfying external demand with the following features:

  • Service Level is 95% (NSP)

  • Forecasts are moderately good

The figure describes the Multi-stage dilemma.

Assume that for the internal warehouse (WH) staging inventory, holding cost is half as expensive as customer-facing (CF).

How should we allocate safety stock buffers across the stages?

  • As much as possible to WH?

  • Shared, more to WH?

  • Shared, more to CF?

  • As much as possible to CF?

  • Half and half?

Different stages should optimally share the risk to:

  • Propagate demand from downstream to upstream

  • Model the impact of the Internal Service Level (ISL) decision of the upstream stage at the downstream stage

Capturing Interactions Between Stages

In multi-stage supply chains, stages are linked together. Orders placed by the customer-facing node create demand for product upstream. The upstream service level has an impact on the ability to meet the service level downstream.

The figure describes the Interactions Between Stages.

Optimal Internal Service Levels Versus Holding Costs

The table in the following figure gives an overview of the optimal internal fill rates as a dependency of different holding costs.

The figure describes the Service Levels and Holding Costs.

Multi-stage Logic for Complex Supply Chains

Simple supply chains can be optimized easily by simple search. For the more complex supply chains that we usually find in practice, sophisticated multi-stage logic is required to efficiently determine optimal solutions. SAP tools provide the appropriate optimization techniques.

The figure describes the Complex Supply Chain.

Safety Stock Buffers for All Forms of Uncertainty in the Supply Chain

Summing up, the following figure provides an overview of the different safety stock buffers for the different kinds of uncertainty that we find in a supply chain.

The figure describes the Safety Stock Buffers for All Forms of Uncertainty in the Supply Chain.

Run a Global (Multistage) Inventory Optimization Operator

Business Example

As inventory planner you would like to run the global (multistage) inventory optimization operator. To do that we will work during this exercise with the connection IBP400 within the Planning Area IBP400.

Prerequisites

Prerequisities

  • Define a Planning Area where this planning should be executed, for example, IBP400.
  • Master and key figure data loaded into the planning area.
  • Template used for this exercise: IO 220 Review Results
  • For this exercise we will work with a planning filter FG1##. Therefore exercise Create a Planning View based on a template is a prerequisite for this exercise.
  • For this exercise the inventory profile created upfront is needed, therefore exercise Set Up an Inventory Profile is a prerequisite.
  • The exercise for forecast error calculation, the supply lead time calculation and the validation of other inputs are prerequisites.

Note

Manual inputs are only feasible in the current and in the future periods. Calculations will not affect the past in the planning horizon

Steps

  1. Create a planning view favorite named IO 220 Review Results ## based on the template IO 220 Review results.

    1. Log on into the SAP IBP, add-in for Microsoft Excel.

    2. In the Planning View area, choose New View → From Template....

    3. Choose IO 220 Review Results as Template.

    4. Add your planning filter FG1## as workbook filter.

    5. Choose Versions & Scenarios.

    6. Select Base Version under Versions and Baseline under Scenarios.

    7. Under Planning Scope, select all sub-networks also called Planning Units that correspond to your group number ##. For example, learners with ## will have all Planning Units containing ##.

    8. Choose the

      Quantities tab.

    9. Choose OK to close the Create New Planning View from Template dialog box.

  2. Go to the Quantities tab to verify that there are no values in this sheet

    1. Verify that you are on the tab Quantities, filtered by your group number ## to display your products.

    2. Verify that Recommended Safety Stock does not contain values.

  3. Go to the Application Jobs group to run the algorithm

    1. In the Application Jobs area, choose Run.

    2. Choose Inventory Planning (Advanced).

    3. Select Global (Multi-Stage) Inventory Optimization as Function under Inventory Planning Profile - ##.

    4. Select Scenarios: Baseline, Versions: Base Version.

    5. Verify, that as Subnetwork 3 ## is selected.

    6. On the Filter tab of the Run Inventory Planning (Advanced) Once dialog box, you will find your planning filter FG1##.

    7. Select Run.

    8. Select as Reason Code: Inventory and add a comment Running the global multistage inventory optimization operator if a reason code is requested.

    9. Click on Run and click on OK to confirm that you have scheduled a job.

    10. Watch the progress in the dialog box. When the run is finished, choose Navigate to Status to view the list of logs. Close to return to planning view.

    11. When the status is Finished, close the window Inventory Optimization - Status and refresh the planning view.

    12. Go to the current period.

    13. Verify that Recommended Safety Stock now do contain results at least for some combinations and periods. If the Standard Unit Cost key figure appears blank, ask the instructor to reload the exchange rate.

    14. Compare the different results from one Location to the other one.

Result

Results

Now you have seen how you can run the Global multistage Inventory Optimization operator and you have analyzed the effects of the inputs over the results.