Comparison of the Capacity Leveling Methods (SAP Library

Comparison of the Capacity Leveling Methods

The following three maincapacity leveling methods are available in Supply Network Planning (SNP):

· Heuristic-based capacity leveling

· Optimization-based capacity leveling

· Your own method connected using the Business Add-In /SAPAPO/SNP_CAP

Which method you should choose depends on the individual circumstances. However, the main features of the two methods provided by SAP are described below to assist you in deciding which to use.

Heuristic-Based Capacity Leveling

Features

Heuristic-based capacity leveling starts from the start or end of the planning horizon depending on the scheduling direction you chose (forward or backward), and compares the resource capacity load in each period with the required load that you defined. If the system detects a resource overload, it first selects all the activities or orders that are causing the overload in the period concerned. The system then sorts these orders according to the priority you specified and, in turn, moves orders or partial order quantities into later or earlier periods until the maximum resource capacity level has been achieved. During forward scheduling, the system moves the orders into the future so that the first activity that uses the resource to be leveled starts after the period with the overload. During backward scheduling, the system moves the orders into the past so that the final activity that uses the resource to be leveled is completed before the start of the period with the overload.

When moving orders, the system takes into account the lot size values and rounding values defined in master data (for information about constraints, see Consideration of Lot Sizes and Other Master Data).

Performance

The runtime of heuristic-based capacity leveling depends on the number of orders to be processed and the number of periods. Runtime is influenced by the following factors in particular:

· Number of products at the resource

· Ratio of the lot size to the total quantity

· Extent of the resource overload

· Choice of periods (day, week, month)

· Length of the planning horizon

Optimization-Based Capacity Leveling

Features

Optimization-based capacity leveling uses the SNP optimizer to clear resource overloads. The system actions are as follows:

Der Optimierer ermittelt zunächst alle Produktionsprozessmodelle (...

...

1. The optimizer first determines all the production process models (PPMs) or production data structures (PDS) and transportation lanes that use the resource to be leveled. It then determines all the location products that belong to this master data plus the relevant master data and transaction data for these products.

2. The optimizer generates a special optimization problem for capacity leveling based on the data determined and the settings made by the user (scheduling direction, for instance). The optimizer automatically sets the costs that are taken into account for this problem (costs for storage, delay, or non-delivery for instance). It ignores any costs defined by the user. The costs are only used to control capacity leveling. They have no business significance.

3. The optimizer solves the optimization problem generated. To do this, it first deletes all orders and stock transfers for the resource to be leveled and then completely replans it. If alternative PPMs, PDS, or transportation lanes are available, it bases its selection on procurement priority, if possible.

The optimizer makes sure that the resource to be leveled is not overloaded and takes care not to move the planned orders and stock transfers too far beyond or prior to the original receipt due dates. It prefers to create receipts that are too early (resulting in the creation of stock on hand) rather than too late (resulting in the creation of shortfall quantities), if possible.

4. The optimizer creates new orders and stock transfers. It then also creates the orders and stock transfers that it was initially unable to schedule due to the limited resource capacity. In backward scheduling, it creates the orders and stock transfers for the original receipt due date; in forward scheduling, it creates them in the last possible period of the planning horizon.

Performance

The following factors have a significant influence on the runtime of optimization-based capacity leveling:

· The length of the planning horizon and the number of periods within it

· The number of alternative transportation lanes and PPMs or PDS

· The number of products for which there are planned orders or stock transfers

The following factors do not have a significant influence on the runtime of optimization-based capacity leveling:

· The number of planned orders or stock transfers for a product within a period

· Lot sizes

· Extent of the resource overload

Comparison of the Methods

The features of the two methods are compared in the following table:

	Heuristic	Optimizer
Type of processing	Order-based processing: · Only orders that are moved are changed (liveCache, R/3) · This is beneficial if there are many orders but only a few that are overloading the resource	Period-based and quantity-based processing: · All existing orders are deleted and recreated after capacity leveling
Resource capacity level	Not always optimal in configurations such as that described in Heuristic-Based Capacity Leveling: Examples	Almost always optimal
Changes to plan	Few	Completely new plan
Priority rule observance	· Order size · Product priority (sorting in ascending or descending order)	No
Overloads moved during backward scheduling	Always	Only if there is sufficient capacity free in the target period
Location of receipt and issue	The receipt has to be within the planning horizon and outside of the SNP production horizon or stock transfer horizon.	The receipt and issue have to be within the planning horizon and outside of the SNP production or stock transfer horizon.
Bucket offset	Is ignored (see period factor)	· The bucket offset for PPMs or PDS is always 1 · The bucket offset for transportation lanes is always 1
Period factor	The period factor of the PPM/PDS or transportation lane is taken into account (if not defined, the period factor of the location product is taken into account)	Is ignored (see bucket offset)