The XYZ company manufactures its product out of 5 different plants in the United States. Each plant manufacturers the same exact product. After manufacturing the product, it is shipped to their distribution center which then distributes them to the various retail centers for sale to the end customers.
Rochester, NY Plant
Philadelphia, PA Distribution Center
Serves the North East US
Altanta, GA Plant
Altanta, GA Distribution Center
Serves the East Coast US
Houston, TX Plant
Dallas, TX Distribution Center
Serves the Southern US
Minneapolis, MN Plant
Minneapolis, MN Distribution Center
Serves the Central and Mid West US
San Diego, CA Plant
Los Angeles, CA Distribution Center
Serves the Western US
By way of example in the North East, the process is as follows: In the North East US, the Rochester plant manufactures the product based on the demand in the North East and then ships it to the Philadelphia Distribution Center. From this center in Philadelphia, it is then shipped to all the stores in the North East.
All the plants and corresponding distribution centers work in the same fashion.
Since the products are identical from each plant, in theory they can close 4 of the plants and have only one do all the manufacturing. However there are cost implications from shipping the product between a plant and a far-away distribution center. There are also a variety of other implications. As such, the current process is that all the products required in a region of the US comes from the plant within that region. Outside of their normal manufacturing production runs, each month the company recieves a recurring order for 10 custom made products. To manufacture these ten products requires processing on two different machines.
Minutes Required By Each Product on Each Machine
| Product |
First Machine |
Second Machine |
| #1 |
28 |
31 |
| #2 |
41 |
29 |
| #3 |
17 |
49 |
| #4 |
51 |
21 |
| #5 |
39 |
55 |
| #6 |
26 |
30 |
| #7 |
42 |
32 |
| #8 |
15 |
47 |
| #9 |
50 |
22 |
| #10 |
40 |
58 |
The short cut approach to causal based forecasting
Step 1 Create 13 variables (1 trend and 12 months)
Step 2 Run the regression and produce the model (run it without an intercept)
Step 3 Graph the current model
Step 4 Using any momentum based forecasting you think is appropriate, forecast the error in the prediction
Then
1. create a historical sales from march 2015 until march 2017 (one or two years)
1. forecast using regression or momemtum
2. Data must be reasonable
3. make recommendation based on production/distribution and quality controls
finished QC is attached
Attachment:- QC.zip