Multistage models allow the internal service level of the upstream stage to be modeled at the downstream stage, and ensure that service level targets are met. This figure shows a multistage cooperative example.
Multi-stage Application via an Example
Objective
After completing this lesson, you will be able to understand the multi-stage inventory optimization capabilities through an example.
Multistage Example
Two-Stage Example: Internal Service Level
Impact on Safety Stock Costs
If we plot safety stock costs against the internal service level, we would like to find an optimal level, where we optimize both variables, such as the optimal internal service level that results in the lowest safety stock holding costs, 72% in this situation. Through SAP IBP for inventory, optimal internal service levels are calculated automatically for every internal product location.
Example: Optimal
Assume an ISL of nearly 100%. Which values in our supply chain would need to change, and would change, to derive the optimal ISL of 72%? The figure provides the answer.
Optimal Risk Sharing Across Stages
Optimal Risk Sharing Across Stages
- A very high internal service level will push all units covering all open backlogs, however, that will not lead to optimal costs (ISL = 99,99%).
- While very low internal service levels will cause higher backlog levels with non-optimal costs (50%).
- That means, a trade-off is needed when balancing backlog levels, holding costs, safety stock, and target inventory position at all positions simultaneously (ISL= 71%).
At Internal Node: Quantifying Unmet Demand
If we consider 100 periods, we can simulate on-hand and order estimations to verify if the demand is met completely or not. At the end of the exercise, we can estimate the average for key figures as on-hand, orders, demand, and unmet demand. In addition, we can estimate the standard deviation to construct a distribution.
The algorithms within SAP Integrated Business Planning (IBP) for Supply Chain estimate the average and the standard deviation as well to calculate the met demand factor.
At CF Node: Impact of Backlog on Safety Stock
At CF Node: Impact of Backlog on Safety Stock
- CF node needs safety stock to hedge against demand variability.
- Additional safety stock is needed to hedge against backlog (service variability).
- Variability pooling between demand variability and service variability reduces safety stock requirements.
The figure shows how the Safety Stock can be calculated considering the backlog.
If all other inputs are constant, for example, service levels:
- Incoming backlog mean will linearly increase safety stock.
- Incoming backlog standard deviation will exponentially increase safety stock.
Drivers of Backlog
Simplifying, there are several drivers of backlog:
- Internal demand characteristics (mean, standard deviation)
- Internal service level
- Exposure (PBR+ LT)
- Lead time variability
Three or More Stages
What if we consider three or more stages within the supply chain? How will the backlog impact the safety stock? The following figure exemplifies the answers to these questions:
Other Drivers of Supply Variability
What if the lead time varies overtime at each station? How high can the changes on the backlog be now that initial conditions are different? The following figures address these two questions, providing insights:
Multiple Demand Streams
If we consider a supply chain with multiple demand streams, we can analyze the changes within the flow of goods and information:
Summary
