Talk:Safety stock/Archives/2012

Use of Safety Stock Calculation for Tank Headroom
If the safety stock calculation is being used for liquid in a tank, can the same formula be used to calculate the required planning headroom for the tank to ensure that, even if demand is less than forecast, the tank can always (usually) fit the contents of a planned delivery on arrival? Samwardill (talk) 14:16, 20 April 2010 (UTC)

Service Level Calculation
Is the "1. Z: NORMSINV(Service level), for example Z=1.96 for a 95% service level" correct? My math puts NORMSINV(0.95) = ~1.64. -DL —Preceding unsigned comment added by Dwlarson (talk • contribs) 10:04, 9 September 2009 (UTC)


 * You are right that there is a problem here. I think the change made by 24.222.2.222 on 25-August-2009 was not right; as a result the example is now messed up. Feel free to correct it. Encyclops (talk) 02:29, 10 September 2009 (UTC)

Older Notes
I have just removed a big pile of commercial links toward consulting blogs. The formula writing is still very ugly and would need a deep refactor.--Joannes Vermorel (talk) 16:28, 8 April 2009 (UTC)

Cancelled the redirect on this to Stock Optimization as there was no decent def in that article as to what safety stock was - be interested in adding reference to specific algorythms that are used to calculate safety stock if anyone would like to participate Shanzu 15:37, 19 June 2007 (UTC)

I have just finished to write a guide "Calculate safety stocks with sales forecasting" at http://www.lokad.com/calculate-safety-stocks-with-sales-forecasting.ashx The guide actually precisely documents the algorithms being used, as Shanzu suggests. --Joannes Vermorel (talk) 14:32, 26 November 2007 (UTC)

For the remark at the bottom about the consistency of units : Yes, it is consistent when you keep in mind that the variation of demand is measured over 1 period (say 1 week), then it is estimated over n periods by d(N) = d(1)*sqrt(n) (student distribution or something). When measured over n periods, the variation increases by sqrt(n).

so the variation of demand OVER THE LEAD TIME is not just "Standard Deviation of Demand" but "Standard Deviation of Demand * sqrt(Avg. Lead Time)". Then the units in the formula are consistent. —Preceding unsigned comment added by 195.6.125.226 (talk) 09:05, 8 January 2009 (UTC)

Dimension error in Safety Stock Calculation
The stock must have units of items (that is, things counted).

The formula has the following units:

SQRT [ ( time * items**2 ) / time**2 and ( item**2 / time**2 ) * time**2 ]

which gives

SQRT[ items**2 / time ] and SQRT [ items**2 ]

which gives

items / SQRT [ time ] and items

Thus the formula has inconsistent dimensions i.e. the two terms do not have the same dimensions. The formula cannot be correct.

The safety stock formula is widely reported and I can find only one other comment on the fact that it is dimensionally inconsistent and so must be incorrect.

Can anyone resolve this problem? Carleysmith (talk) 09:55, 9 May 2012 (UTC)

Agreed, I did post a dimensionally consistent formula at http://www.lokad.com/calculate-safety-stocks-with-sales-forecasting.ashx However, I did not invent it, this formula has been around for decades in supply chain manuals. I would not trust too much the formula listed presently on the main article. --Joannes Vermorel (talk) 11:31, 24 May 2012 (UTC)