Foreacast Consumption in Time Series

[former_member592774]

Hi Experts,

I am trying to perform the forecast consumption in IBP Time Series Supply , I am having some issues in values of some standard output KF.

I have created a forecast consumption profile with 4 periods backward and 1 period forward consumption.

The level of Consumption is Product - Location.

I am trying to use the input demand/forecast KF from a KF which is equal to the standard Dependent Demand KF ( which is at Prod-Loc level)

Sales Order Input is also at same level.

When I run the consumption independently as separate S&OP operator the results are good.

But when I try to run the forecast consumption as preliminary step of Heuristic , I see that Dependent Demand KFs are being cleared and hence the Customer Dependent Demand/Customer Receipt is also not calculated.

Could you please help understanding the logic here?

Thanks in advance.

Regards,

Anurag

[mkorndoerfer · Accepted solution]

Hi Anurag,

It could happen that when Heuristic is run, during the Demand Propagation step, there is no demand in downstream direction to be propagated upstream. That is, from where Dependent Demand of Location LOC1 is coming from? From Customer or another Location?
Make sure there is demand downstream.

Also, can you please confirm which Key Figures are the Input Forecast, Input Sales Order, Output Open Forecast and Output Total Demand?

Best regards,
Matheus

[former_member592774 · Accepted solution]

Hi Matheus,

Thanks for your response.

Dependent Demand on LOC1 is coming from Customer. So when we just run the heuristics the Dependent Customer Demand at LOC1 is total Dependent Demand.

The Input Forecast is my case is :Z Forecast ( which in KF calc = Dependent Demand KF)

Input Sales Order = Sales Order KF

Output Open Forecast = Open Forecast KF

Output Total Demand = Output Forecast Consumption

The main question here is that when forecast consumption is run together with heuristic , Net Demand is calculated correctly but it clears Dependent Demand and hence all the Customer Dependent Demand and Customer Receipts.

What I expect is that it computes the Net Demand and but Keeps/updates the Dependent Demand/Independent Demand so that we can show the customer receipts.

Regards,

Anurag

[mkorndoerfer · Accepted solution]

Hi Anurag,

I'd suggest you create an incident under component SCM-IBP-SUP-ALG so we could analyze all the supply chain and Key Figures. I think it's more feasible than continue working through this channel.

As soon as resolution is provided, you may update it in here.

Best regards,
Matheus

[gurucharanscm]

Hi Anurag - you have LOC1 in the mid of your supply chain (at least it's not the destination), so it can be presumed as PLANT >> LOC2 >> LOC1 >> CUST. And you do the fcst consumption at LOC1 (by assigning a custom KF as your FCST key-figure at LOC1).

1. Fcst consumption is allowed to run before Heuristic and it does it’s job right. The net demand is shown rightly.
2. Later, Heuristic will not consider any INDEPENDENT DEMAND at the location where the fcst consumption is carried out (i.e. at LOC1) as that is substituted by Output total demand volumes (net demand).
3. You don't have anything as INDEPENDENT DEMAND at LOC1, so that's fine.
4. Heuristic then has to carry out the algorithm run. It doesn't do a DELTA run, it is always FULL.
5. By FULL, I mean, it deletes all the existing receipts/supplies and recretes everything afresh across your supply chain.
6. So, the previously created outbound/dependent customer demand gets deleted as part of point 4).
7. The planning now happens considering the net demand at the location where forecast is consumed and doesn't consider the downstream location - I am presuming this because of point 8).
8. If it does consider the demands at a downstream location CUST to the location where you carried out your consumption (LOC1), your consumption quantities will be of no use - so I presume it's designed that way.
9. I presume you will end up having more customer receipts if you choose to run the consumption profile independently and then again carry out the Heuristic run independently - because, it now has dependent customer demand at CUST1 + your big sales order volume at LOC1.

I suggest to show up the sales order volumes in IBP at PROD/CUST location (taking the SHIP-TO from the sales order) and have the consumption profile run at the PROD/CUST level.

I would prefer other colleagues to put their view point as well.

Thanks,

Regards - Guru

[former_member592774]

Thanks Guru for detailed analysis and inputs.

I agree that I have CUST as destination in my supply chain that's why I wanted to have the customer receipts calculated even if I do forecast consumption at DC level (which of course serves the customer).

I agree to me it looks like when we run heuristics to consume the consumption results , it take the inputs as Net demand at DC level to start and propagate the demand/receipts further to Plant and Suppliers.It ignores the Customers then hence customer dependent demand and the receipts are deleted.

Thanks for your suggestion.

Regards,

Anurag.