On large assembly lines, setup is often problematic. Parts are  brought in for a new job and remaining parts taken away for the just-completed job. Such lines often need re-balancing with different stations and task assignments. Then, there is the initial startup as everyone gets accustomed to the new configuration. If parts do not fit, or cannot be found, there is more disturbance. When, at last, the line is humming, nobody wants to disturb it with another changeover. Such is the case with our hypothetical assembly line.

And so, once setup, the line runs for a week on the first product, 1-GRN. At the end of a week, 40 units are complete even though the customers only need 10 units. The remaining 30 units must go into inventory to carry the customer demand for the next three weeks. The figures below illustrate.

After another frenetic changeover to product 2-YEL, the line runs for 2.0 weeks since the demand for this part is higher. Eighty units are built. Twenty of these units supply customer demand during the run. The remaining 60 go into inventory.

Another changeover and, for a week, the line runs 3-RED. Forty units are made. Ten for customers and the remaining 30 for inventory.

The three suppliers for the upper covers of, respectively, 1-GRN, 2-YEL and 3-RED, see high demand for one or two weeks and then no demand for 2-3 weeks. The charts show this erratic demand.

One way to deal with this intermittent upstream demand is to size upstream production for the overall average demand for each part and then build inventory between production runs. This, however, creates a host of inventory control, scheduling and quality problems.

The Mixed Model Assembly line addresses this situation.