How system is calculting propogated demand in multi-stage inventory optimization model?

[former_member590783]

Hi,

I have multi-stage inventory optimization.

I have sum up IO forecast from lover level node. Propagated demand at higher node at approximately same as summation of IO forecast at lover level but not exactly same.

Any information how exactly propogated demand is getting calculated will be appreciated.

Thanks

Amit

[former_member185660 · Accepted solution]

Hi Amit,

In SAP IBP for inventory, propagated demand mean depends on the type of location, e.g, customer facing,single sourcing up-stream, multi-sourcing up-stream, hybrid.

For a customer facing location, propagated demand mean = forecast.

For a single-sourcing up-stream location LOC2 --> LOC1,

propagated demand @ LOC2 = propagated demand mean (future period) @ LOC1 + recommended safety stock (future period) @ LOC1 - recommended safety stock (current period) @ LOC1.

The future period values are defined by PBR.

For a multi-sourcing up-stream location, the logic is the same as the single-sourcing up-stream location, except the formula requires to consider sourcing ratios. For example, for LOC3 sourcing both LOC2 and LOC1,

propagated demand @ LOC3 = (propagated demand mean (future period) @ LOC1 + recommended safety stock (future period) @ LOC1 - recommended safety stock (current period) @ LOC1) * sourcing ratio (LOC3->LOC1) + (propagated demand mean (future period) @ LOC2 + recommended safety stock (future period) @ LOC2 - recommended safety stock (current period) @ LOC2) * sourcing ratio (LOC3->LOC2)

I hope this brief explanation helps.

[former_member590783 · Accepted solution]

Hi Alexis,

I come across one mysterious situation today. In alternate week demand is getting in propagated demand at supplying node.

Like for week of 9/10, 9/24, 10/8 demand getting propagated from lower nodes but not for 9/17, 10/1, 10/15.

propagated-demand.png

Amit Agrawal

[former_member185660 · Accepted solution]

Hi Amit,

Do you have forecast inputs for those weeks?

Are you modeling time varying sourcing ratios? If so, can you check the sourcing ratios for the supplying node for those weeks?

Regards,

Alexis Lozada

[moggiemx79]

This is the formula I translated from what Alexis Lozada wrote, just needed to be a little bit more clear about the periods for which you calculate Propagated Demand (including LT and PBR):

Thanks,

Héctor

[former_member589700]

Hi Alexis,

For hybrid node, how it is calculated?

[former_member590783]

Thanks Alexis.

[pluthaman]

Hi Amit, Usually if your KF calculation @REQUEST is all correct - the sum of the lower level nodes (with decimals - usually decimals aren't visible in excel unless you put your cursor in that cell) should be aggregated level value.

Thanks,

Uthaman