Designing the Mixed-Product Sequence

Why it matters

The mixed-product sequence is not the same thing as the decision of how many units of which product to produce. Even if the mix is right, load can build up at a certain station when heavy jobs come one after another. When light jobs come in between, the same total load can be distributed more evenly. To see this difference, short time windows are examined instead of only the daily average. Sequence design does not magically reduce the total time; it changes the distribution of the existing load over time. That makes visible at which point the limited waiting area, worker support or delivery window is strained. The calculation must be done while keeping the same product quantities.

Average load and load peak are separate results. If a sequence's average equals the takt time, there may be no excess load in total; but the work of a single product can exceed the takt window. Whether this overrun can be safely completed in the window of the next light product depends on the physical layout. If there is a strict transfer limit, carrying work over may already be forbidden. If there is a flexible station or a defined waiting arrangement, limited carryover may be allowed. This condition is written before the calculation. The overrun is not closed by assuming that the worker will speed up or that an invisible helper will come. Numerical acceptability does not replace a safe method and priority relationships.

A sequence change also changes the number of setups and the changeover directions. A sequence that spreads load well may require more frequent fixture changes. A sequence that places similar products side by side reduces setup but can increase the station load peak. For this reason the sequence is not chosen by a single criterion. Processing, setup, the time products become usable and material presentation are evaluated together. The total of a candidate with a missing changeover time is not written as if it were precisely calculated. Whichever part of the decision the unknown information limits is stated explicitly. Design requires being able to hold back a sequence for which there is insufficient evidence, as much as choosing the sequence that looks suitable on the numbers.

The relationship between the end and the beginning of a repeated sequence must not be forgotten. In one schedule the last product may be light and the first heavy; in another sequence two heavy products may join at the boundary. For this reason, watching only the inside of one repeat hides some peak values. The steady repeat condition, whether work is carried over at the start, and the end-of-day cleaning load are stated separately. Even if all parts are completed at the end of the day, a delivery within the day may have been missed. By comparing two sequences of the same mix under equal data, this lesson establishes the relationship between the average, the short-window load and the carried-over work.

Formula