CTM run in APO

Question

Dear All,
I am learning APO.I can do the
DP & SNP (Network ,Multilevel)
heuristics Runs on APO 3.1 very well.
I would like to know how to Run CTM ? What is CTM? and most important what are the settings required for CTM run in APO?
Can anybody help & guide me.
I am awaiting for your valuable answer.Many thanks in advance.
Warm Regards,
Vishwesh Rajguru

Answer 1

Waht is CTM
Capable-to-Match (CTM) functions match a large set of
prioritized customer demands and forecasts to a set of
categorized supplies considering the current production
capacities and transportation capabilities in a multistage
production environment, as in the semiconductor
industry. CTM is based on constraint-based propagation
techniques and goal-oriented programming.
Preprocessing tools provide a set of categorized supplies
and prioritized demands as input to the CTM Engine,
which then performs a fast check of production
capacities while considering transportation capabilities.
Prioritize by Demand
CTM prioritizes demands, such as customer orders
or forecasts from SAP APO Demand Planning, based
on such characteristics as customer priority, location,
and product priority.
Supply Categorization
CTM performs supply categorization based on the
available supply and user-defined inventory limits, such
as amount of inventory or planned deliveries.
CTM Engine
The CTM engine delivers the functions you need to
control supply and demand. The CTM engine process
can be divided into two steps. In the first step, the
CTM engine builds the CTM application model using
the specified supply network model. In a second phase,
the CTM engine matches demand to supply considering
production capacities and transportation capabilities
and using constraint-based propagation. CTM searches
before production and after production.
In before production searches (or most to least), the
system first searches through all supply categories for
finished or semifinished products before it resorts to
ordering production for that item. The search is
conducted top down. The system first searches the
production process model for finished products, then
intermediate products, then raw materials.
In after production searches (or least to most), the
system searches for categories after an attempt is made
to produce the item.
Time-Dependent Production Processes
You can define different production processes,
including product flow, for different time intervals. The
CTM engine chooses only production processes that
are valid at the representative time.
Model Alternative Components
You can implicitly model alternative components by
annotating a list of alternatives within the bill of
material, or you can model alternative components
explicitly by defining a set of product substitutions for
a given product.
Identification of Raw Material
A special product category identifies material that is
available in an infinite quantity. But for the consumption
of this product, a purchase requisition must be
created.
Process Flows Based on Demand
You must select a specific process flow based on the
set of characteristics of a given demand. For a planning
process that covers several physical plants, you must
be able to restrict the production and distribution
network to a set of allowed plants. In addition, the
CTM engine can handle the plant substitution rules of
SAP APO’s Global ATP component.
Product Substitution Rules
CTM considers a set of demand characteristics for each
product and a prioritized list of substitute products
using Global ATP rules.
Customization/Settings
Global customizing settings
The package size is the key parameter that can improve the performance if you are using the asynchronous liveCache (LC) update.
In asynchronous update the system creates the orders in the LC after planning the specified number of demands. At the same time, the system continues with the CTM planning for the other demands.
The performance depends on the amount of planned orders that are created per demand.
If you set up a package size of 1000 demands, this means that each 1000 demands CTM writes the results to the LC. If there are only 1000 planned orders created for fulfilling these 1000 demands then the package size is too small for an asynchronous run with a good performance because the CTM run needs to stop each 1000 planned orders for triggering the LC writing. This increases then the run time. In case there are 100.000 planned orders created then the package size is too big because the CTM run for creating these 100.000 planned orders needs probably longer than the parallel writing of the results to the LC.
So the optimal package size makes the planning and writing to LC almost parallel.
Apart from the package size of the asynchronous LC writing, there are three more areas where packages must be defined.
- Package Size for Creating Orders. The value indicates how many orders can be created in the LC at one time.
- Package Size for Creating Pegging Relationships. The value indicates how many pegging relationships can be created in the LC at one time.
- Package Size for Order Selection. The value indicates for how many location products a LC request can be started at one time (package size for reading orders from LC in a way that CTM can use them for planning).
It is difficult to give recommendations regarding these package sizes because one needs to play around in a specific situation to improve the performance. So consider the following values as an orientation.
Regarding the package size of the Order Creation a value smaller than 1000 can be recommended (default value is 500 orders), but also a value up to 5000 can be efficient. In case of performance issues it is recommended to run CTM with different package sizes : 500, 1000, 2500, 5000 and see whether one of these settings can improve the performance.
Regarding the package size for Creating Pegging Relationships a value bigger than 1000 is recommended. The default value is 5000 pegging relationships.
The setting for the Order Selection is based on location-products, which implies that if the setting is currently 5000, CTM reads the orders for the first 5000 location-products and sends them in one package from the LC. If the average amount of orders per location-products combination is 10 (each location-product has in average 10 orders) the result is 50.000 orders transferred in the package.
As this setting depends on the amount of orders existing per location-product, one needs to play again with different settings in case of performance issues. Generally it is recommended to have a value smaller than 10.000.
3) Planning profile (business improvements)
Apart from the customizing settings, the performance depends on many other criteria like the Master Data Selection, Order Selection, Selected Demands, Planning Period, Product-Location restrictions.
If you want to improve the performance you should create a master data selection which contains only the objects in the model which are actively planned by the CTM run.
You can check whether it is possible to divide your supply chain in separate independent model parts. This covers a big performance improvement potential because the independent parts of the chain can be planned in parallel.
Then the order selection, especially the size and data requirements, is one of the major areas where you can improve the performance.
During the deletion of orders in a CTM planning run, several checks that are not always required are run. The parameter 'FAST_DEL' skips some checks of the normal planning run to improve the planning run time. The parameter is activated if you set Value1 of the planning parameter to X. In this case, no information is analysed on pegging, input or output nodes during order selection. Therefore, the parameter can only be reasonably used with the planning mode "Regenerative planning" (Replan all Orders) and the deletion mode "All unfirmed orders" (Delete all Orders that are not firmed). Note that in this case, only the status and the type of the order are analyzed for the deletion decision. This means that all planned orders or purchase requisitions are deleted, if they were not firmed manually (Output fix) and contain at least one location product that is contained in the CTM master data selection. This parameter must not be used if you use a subcontracting scenario. In this case, inconsistencies can occur between the data in APO and the R/3 System because subcontracting purchase requisitions are deleted but not the subcontractor planned order at the subcontractor side.
Notice that since release SCM 4.0 the parameter 'FAST_DEL' is obsolete and has been replaced with the planning step (Technical Settings tab) End Planning Run after Oder Selection in combination with Do Not Check Order Details.
CTM run time depends on the number of demands. Consequently if your business need for CTM does not require all demands simultaneously, you can perform CTM runs using a demand selection. You can cut down on run time if you can first reduce the amount of demands considered in the planning run.
Another factor to take into account is the planning period set for the CTM profile. You can implement a horizon in the work area which includes only those demands that can be planned during the horizon of the CTM run. That is, exclude any periods containing demands which cannot be planned.
Consider also whether you can meet your business requirements by restricting your demands by product-location. You can experience long run times because in a large supply chain you do not restrict the demands by product-location combination.