Difference Between Statistical Forecasting, Constrained Forecasting and Unsconstrained Forecasting

[#1]

Hi,

Can any one pls tell me the differences between Statistical Forecasting, Constrained Forecasting and Unsconstrained Forecasting with an example.

Regards,

Kumar.

[rajkj]

Kumar,

Consider a make-to-stock example. A packaged foods manufacturer supplies frozen meat products to retailers. For brevity, assume a single product and production capacity is limited by the raw material i.e. right animal (or bird) with respect age, weight, health, etc. There is a confirmed supply of 2500 kg of meat per month.

The sales demand (cleansed history) for past 3 months are as follows: (Comes from the sales history)

Category	April	May	June
Cleansed Sales history	2400	2600	2500

Considering a simple moving average method to generate statistical forecast and a simple promotion to take advantage of the summer month's seasonal demand.

Category	July	Aug	Sept
Stat. Forecast	2500	2533	2511
Promotion	600	1000	500
Total unconstrained forecast	3100	3533	3011

Since the available supply is only 2500 kg per month, the manufacturer has to look outside to meet the demand spikes (especially due to promotion). Assume the following confirmed capacity.

Confirmed Capacity	July	Aug	Sept
In-house production	2500	2500	2500
External procurement	450	600	511
Total confirmed capacity	2950	3100	3011

Based on meat availability constraint, the total confirmed capacity shown above will be our constrained forecast that can be met by the production.

Category	July	Aug	Sept
Unconstrained Forecast	3100	3533	3011
Constrained Forecast	2950	3100	3011
Difference	150	433	0

Since the difference is significant in July and Aug months, the manufacturer has to rework on the planned promotion for these two months.

I hope this simple example helps you to understand the business perspective of statistical, unconstrained, and constrained forecasts.

Thanks,

Rajesh