Setting Up Fixed and Variable Production Costs for SNP and PP/DS

Notice that the costs for both storage and safety stock penalty are declared as a daily value right in the Location Product Master interface.

Background

A common need is to setup production costs in the system for production so they can be picked up by the optimizer. In this example I will be using the PPM or Production Process Model.

How Costs Are Used by the Optimizer

Both fixed and variable costs can be setup in the PPM. Secondly, these same costs can be setup for either single level or multi level. The single level applies to SNP and the multi level applies to PP/DS. Different PPMs can be setup with different costs which are then selected by the optimizer. Most companies that I have worked with set a single cost per line in their factory, which is then represented by the PPM. So if there are 4  lines in a factory, and 10 products made per line, then there would be 40 PPMs, each with their own cost. However, this is just one way to perform the setup, there are many ways to perform a setup.

What the Costs Mean

The majority of costs that are used by the optimizer to calculate a plan output are variable costs. This applies to the cost of a missed demand, the costs of violating safety stock, storage costs and transportation costs. Therefore the cost is set per the basic unit of measure in the planning system, and the costs are incurred per day outcome occurs. So for instance, storage costs in SNP are incurred per unit of measure (case, drum, etc..) per day. So if the storage costs are 50 cents per day per drum, and there are 100 drums stored for 10 days, then the storage costs incurred would be $500. This applies to almost all the costs used by the optimizer.

Fixed and Variable Costs

However, production costs offer the exception to this rule. A PPM offers both a fixed and variable cost. So if the variable cost is $40 per barrel and the fixed costs (which could be the setup costs combined with the down time costs) was $1000 and  100 barrels were produced than the cost would be $1000 + (40 x $100 = $4000) or $5000. If I wanted to setup the PPM this way, I would set it up as the following:

The optimizer can compare the storage costs against the production costs, as well as the other three major costs in the SNP (transportation, safety stock, cost of missed demand). For PP/DS these are not the relevant costs for the optimization. Also, there is an overlap with the fixed costs and the Setup Matrix, which is a more sophisticated way of setting the costs and time of moving from one product to another. However, given the long running problems with the Setup Matix, this can be a way of getting around Setup Matrix. However, the functionality of a variable cost per production, which does not consider the next item to be run, is far lower than the hypothetical functionality of the Setup Matrix. But the Setup Matrix functionality is just hypothetical.

How it is Frequently Done in Practice

While most clients I have seen only populate one cost, either variable or fixed, in fact it makes a lot more sense to enter values in both fields. The fixed costs provide a predisposition against making a production order, which is traded off against the storage costs. In fact both the fixed costs and the minimum lot size predisposition the system against creating production orders, so the fixed cost in the PPM and the lot size should be seen as working in conjunction with one another.

Conclusion

The production costs settings are very important and generally not well setup or well thought through. One obvious improvement is to populate both fixed and variable values. They should also be somewhat realistic. There is a strong tendency in cost optimization is to enter completely phony costs, and when this happens the costs become divorced from reality which leads to negative long term implications for advancing the solution. However, attempting to get them to be perfect is also not a very good use of time. This topic is described in this post.

http://www.scmfocus.com/supplyplanning/2011/07/09/what-is-your-supply-planning-optimizer-optimizing/

http://www.scmfocus.com/inventoryoptimizationmultiechelon/2011/05/how-costs-are-really-set-in-cost-optimization-implementations/

Deleting PP/DS Orders with /N/SAPAPO/DELETE_PP_ORDER

Background

APO has a well known transaction called /N/SAPAPO/RLCDEL which will delete SNP orders. Very surprisingly, this is listed on some websites as /N/SAPAPO/RLCDELETE, which is not correct. If you type in /N/SAPAPO/RLCDELETE into the the APO execute box, you will not get anything back. /N/SAPAPO/RLCDEL is covered in this post:

http://www.scmfocus.com/sapplanning/2011/02/02/deleting-transaction-data-from-apo-and-erp/

Deleting orders is often necessary when dialing in APO in order to remove orders that are unwanted. However, deleting PP/DS orders have a different transaction/program. This program must be run from SE38 and is /N/SAPAPO/DELETE_PP_ORDER.

Once the program is typed in, select the green checkbox to run the program. This will bring up the selection screen.

The order category can be selected which can AI or AJ and then running this with the order category selected and any other sub categorizations of say location or product will run the delete job for just that portion of the supply network.

Below, in the control parameters, the Delete Orders must be selected. If the Display Orders selection is only selected, then SAP will simply show you what is to be deleted. The final selection, “Send Deletion” will delete the production orders from SAP ERP as well as from APO.

Sticky Production Planning Orders

However, not all production planning orders can be deleted so easily. When the program above is run, but does not delete planned production orders, it becomes necessary to delete them in ERP. This ERP transaction is CO78.

We will select Deletion flag/ Deletion Indictor. This will bring up the screen below.

Next this screen will come up.

We will select For Individual Selection Screens and then the first selection, and once on the selection screen, the first problem presents itself. That is that the order type listed here does not match the order type in APO. The order type in APO that we are attempting to delete is AI, which is a planned PP/DS order.

The Problems With PP/DS for Process Manufacturing Clients

Background

The software selection mistakes that are made by the large consulting firms never cease to amaze me. There are all types of selections errors that I run into on a frequent basis, however, none of them as bad as the selection of PP/DS for process industry.

The Mismatch

I will say explicitly, PP/DS has no capability to ever have implementation success in a process manufacturing environment. PP/DS lacks the functionality the process manufacturing companies require across the board and the gaps are not even close. The list of incompatibilities with process manufacturing as as follows:

These are serious shortcomings, however, this does not prevent SAP, or large consulting firms from continuing to recommend PP/DS for process accounts.

Understanding the Limitations of PP/DS

PP/DS was designed for discrete manufacturing environments only. This is true because the software has seen limited development effort, and static discrete manufacturing is the simplest manufacturing the plan.  PP/DS’s only can offer basic interfaces such as the Detailed Scheduling Planning Board and the Resource Planning Table.

No Lagging and Overlapping

An absolute requirement for process manufacturing is the ability to lag and overlap operations. That is a second operation needs to start after a certain percentage of a first operation is complete. However, PP/DS does not speak in these terms. It understands scheduling two operations at the same time (SS) or beginning an operation after the first operation as completed (FS). Just this limitation should restrict the application from being selected for process manufacturing, yet it doesn’t.

Inflexible Resource Capacities

Resource capacities must be changed on a frequent basis in process manufacturing. However, PP/DS’s PPM, PDS, and Resources only allow for copies to be made, and for the selection of these alternate objects. However, this creates a significant maintenance issue which many companies are not aware of before they implement PP/DS. In actuality, there has never been a good reason for the existence of PPMs and PDS in the first place. Other scheduling applications keep the BOM, resources and routing separate and relate them to one another. This allows any object to be changed and based upon links or relationships to be updated. SAP has been strongly pushing PDSs over the past several years, but they are even higher maintenance than PPMs. Overall, PPMs and PDSs are simply a bad design that end up taking up a large amount of time and money to support, and take up too much implementation time setting up and even discussing on projects.

Conclusion

The disregard for their client’s implementation success is demonstrated by large consulting companies that continue to propose APO modules that are destined to fail in the environments which they are being recommended. Oftentimes one hears about the relatively low success rate for enterprise implementations, and often in these articles different factors are listed. However, one of the most obvious, blatantly incorrect software selection, never seems to be on that list. It really should be, because bad advice on the appropriate software to use is rampant in the enterprise software industry. No doubt this will be seen as a controversial article. It shouldn’t be. In cases where SAP’s module lack’s the capability to be successful at a company, it should not be recommended.


Resource Variability Capabilities in SCM


Managing Resource Variability

When the topic of resources comes up on projects invariably the question arises how the resource can be made flexible to account for the constantly changing factors in the supply chain.

One way of dealing with constraints that are constantly changing is with the use of capacity variants. Variants allow SCM to modify resource capacity by time of day, by week, or by month. This can help in making the resources “flex” through time to match the reality of the supply chain. Different capacity variants can be used for different planning runs. A second method is using alternate production process models (PPMs) which means using different capacity constraints. Jorg Dickensbach has the following to say on this topic.

In addition to alternative PPMs and PDSs, SCM has the ability to use alternative resources. This allows the system to select from say one machine that is more expensive, but faster vs. a machine which is less expensive but slower. Depending upon the circumstance, either machine can be selected depending upon how it fits the need. - Jorg Dickensbach

Willingness to Maintain the System Master Data?

A second issue which is not very often discussed is how capable companies are in maintaining these constraints. The answer is not very. Most companies I have consulted with basically want the system to maintain itself, and the trend towards outsourcing system maintenance in faraway locations where the individuals are purely technical, and lack business knowledge is only exacerbating the situation. The best systems are those that are self adaptive. I have a video from the CEO Joseph Shamir on this topic from the SCM Focus Inventory Optimization and Multi Echelon site in which I think he expresses very well that clients want in terms of maintainability.

Title: Joseph Shamir of ToolsGroup on What Clients Are Looking For

So we learned from Joseph that demand shaping is one thing that clients are asking for. Also that there is a recognition that many supply chain planning systems end up being difficult to maintain. As a personal comment I think Joseph is absolutely right, and I hold the consulting firms as accountable as the the vendors as many solutions are designed without consideration to whether they will be truly sustainable. Its good to see a vendor with an eye on this issue.

Unfortunately SAP APO/SCM is not there yet. SCM requires a high degree of maintenance in all of its master data.

MDM to the Rescue? (Not Really)

One might ask the question if whether the expense of implementations is actually sucking money away from the maintenance of systems. This could be the topic of a separate article by itself. So while Master Data Management MDM software is selling very well and MDM consultants are doing great, master data is not improving much if at all, (repeated use of the terms “governance” and “taxonomy” do not seem to be magically improving the master data) and the vast majority of MDM projects are simply consulting boondoggles. MDM is known among partners as an excellent way to extract large amounts of money from clients.

See this article on this topic if you are interested in details as to why.

http://www.scmfocus.com/supplychainmasterdata/2010/06/why-software-based-mdm-is-a-consulting-boondoggle/

Secondly, one topic is not discussed with regards to master data maintenance is in many ways getting worse due to outsourcing. Outsourcing of IT is essentially a way to get short term costs reduced, while reducing IT capabilities even more than the cost go down. This means less support for everything and system maintenance is one of them. Its automatic, as soon as outsourcing occurs, the business starts having to perform their own system maintenance. Master data is of course part of this.

Other Methods

Maintenance in SCM can be managed by maintaining Setup Groups and the Setup Matrix in an Excel file and using it for an upload of the Setup Groups to SAP ERP with a CATT via LSMW. Step two is to transfer the Setup Groups to APO with a report, and the third step is to load the Setup Matrix from Excel to APO using the APO loader.

References

http://www.lib.umd.edu/drum/bitstream/1903/7488/4/25813_cov.pdf

Supply Chain Management with APO, Jorg T Dickensbach

http://planettogether.com/capacity-planning.php


PP/DS Optimizer


Benefits of the PP/DS Optimizer

The benefit of the PP/DS can be valuable for reducing sequence dependent setup times, reducing lead times, and to schedule orders according to the priority of the demands.

The downside is that the results are not easy to understand.

Michael Pinedo points out that the objective of scheduling is to minimize the sum of all penalties.

The optimizer uses its own data model.

Pegging constraints and shelf life are used by the optimizer, but adds complexity. The demand at the top level is considered a soft constraint. It is only possible to prioritize demands in the optimizer profile. The optimizer does not take into account alternative PPMs or PDSs.

The optimizer profile contains controls over the following:

  • Horizon and begin
  • Weights of the target function and the optimization algorithm and runtime
  • Order selection
  • Backwards pegging? (one of the most powerful settings in PP/DS)
  • Use of finite resources
  • Prioritization
  • Backward scheduling
  • The target function for the optimizer contains the criteria

Lead time of operations

  • Setup times
  • Setup Costs
  • Average lateness
  • Maximum lateness

Each of these is then weighed. The relative weighs determine the solution.

wpid-optimization-2010-05-5-06-29.jpg
In the optimization profile, the runtime and weightings can be changed interactively before the optimization run. You decide the potential for the quality of the solution by setting the runtime for the optimization. The way in which you weight optimization criteria to obtain good solutions is dependent on the scheduling situation and the organizational goals.

Killing the Optimization

Interestingly, as of 4.0 the optimizer can be stopped and intermediate solutions can be reviewed and used.

Where

The optimizer can be called from the following locations:

  • DS Planning Board
  • Product Planning Table
  • Production Planning Run

Finite or Infinite Production Resources

Resources are considered infinite unless the button “schedule according to the settings in the master data” is set in the order processing view of the optimize profile.

The optimizer considers operations fixed, and pegged orders for external procurement (defined in the optimizer profile). Planned delivery time to purchase requisitions are considered if the flag has been set in order selection of the optimizer profile.

Finite or Infinite Calendar Resources

If a calendar resource is used as handling resource, the goods receipt activities do not consume any capacity and therefore any number of activities for the optimizer to not increase.

Supply Date and Demand Dates

Supply dates are regarded as hard constraints, while demand dates are considered soft constraints. If the bottleneck is at the end of the material flow, it is possible to ignore the supply dates of orders on not selected resources with the setting of “no not consider upstream dependencies” in the optimizer profile.

Prioritization

Demands can be prioritized according to their delivery priority, orders according to their order priority and status.

Complexity

The optimizer in PP/DS has been tested up to 200,000 operations.

Setting Costs

The optimizer in PP/DS is driven by costs. These costs are entered here:

There are both costs for activities, and costs related to delays. The relative magnitude of the costs controls the optimization. The objective function is to minimize costs.

Converting Planned Orders to Production Orders

This is triggered in APO or in SAP ERP. /SAPAPO/RRP7

Triggering the conversion from APO has the advantage that the connection between the deleted planned order and the new production order is considered, so that the production order is matched with the planned order and the operation dates are kept

The conversion of the planned order usually has to take place some time before the scheduled start to provide enough time for the preparation tasks as the printing of the order papers and transport of the components of the resource. This time buffer is modeled by the opening horizon in the PP/DS view of the product master.

Triggering the conversion from APO is only possible for PP/DS orders and not for SNP orders.

References

Jorg Dickensbach

Michael Pinedo


The Implications of the Product Master Priority Code

What Is It?

The priority field sits in the SNP2 tab of the product location master and controls quite a bit of cross module functions in SCM.


In PP/DS

It controls the order by which material is scheduled in PP/DS. Thus for all production orders for a specific date and for a specific production line (priority can only be used to sort among products to be applied to similar resources) the priority helps to sequence the production orders.

In SNP

Used for the deployment optimizer and in capacity leveling. However, since the deployment optimizer is rarely used, it is primarily in effect for capacity leveling.

Conclusion

Overall the priority field is a very simple field that can affect both PP/DS and SNP. I have run into it as used by PP/DS for sequencing.