Skip to main content

Hi,

We are implementing Demand Planner for a client who uses apps-10. We are facing an issue there. its like, based on the forecast models we are getting different forecast outputs. but the problem is for the first forecast month we are getting the desired result but for the rest of the months we are getting either the same value or only increasing values or only decreasing values. but the expectation is to get different values that may fluctuate month by month for the forecasting periods.

For example, if we are using 3 months moving average, in the first forecast period it system gives the forecast value averaging last 3 months demand. for the 2nd forecast period it is expected that the first forecast value should be considered but system is pasting the same value as the first period and all other forecast periods (in our case 12 periods) are also having the same value.

Would appreciate if anyone can help in understanding this logic.

 

Regards-

Nahin

Hi,

When you use the Moving Average forecast model it will always be a straight line (based on the average demands of the past 3 months in your example).

Try switching to a different forecast model such as EWMA Level and Trend which uses the concept of Moving Average but can add weight to the most recent demands or most recent trends.  You can do this directly in the DP Client.

Alternatively, you could try forecast model Bayesian which is a good starter if the customer has not used statistical forecasting before.

I suspect that you may also need to include Seasonal Profiles depending on what forecast profile the customer is expecting to see.

Hi @Richard Owen ,

We are ok with the concept of seasonal profile. but our point is, if its a 3 months moving average and last 3 months demand is 10, 20, 30 then the 1st period forecast should be (10+20+30)/3=20, 2nd period forecast should be (20+30+20)/3= 23.33, 3rd period forecast should be (30+20+23.33)/3= 24.44, 4th period forecast should be(20+23.33+24.44)/3= 22.59 and so on.

that’s what our expectation is but the system shows 20 for all the forecast periods. Need a clarification for this.

 

Regards-

Nahin

From an APICS/ASCM manual - “...the forecast quickly becomes repetitive when forecast data are used in place of actual results; after a few periods, everything is based on forecast data and a moving average or weighted average of the same three numbers results in the same number. It shows how these methods are less useful the further you go into the future.”

Maybe the logic acknowledges this and just uses P1 forecast for the entire horizon.

Same scenario with smaller gaps in actual demand numbers.

 

Hi,

The Moving Average functionality doesn’t work like that in IFS Demand Planner.

The system computes the sum of the historical demands for the past 3 months and divides by 3 (in your example). 

This value is then used as the period forecast value for the entire forecast horizon (12 months in your example).

The IFS Help describes the functionality like this;

Moving Average  

Sets all the periods of the forecast equal to the average of historical demand. The average is based on the number of most recent periods entered in the Moving Average Periods.

Terms & ConditionsCookie settings