The Every Product Every Interval Calculation

Why it matters

The every-product-every-interval approach asks at what interval all the products in a product family can be produced again. The established abbreviation EPEI describes this repeat interval. As the interval gets shorter, products can be renewed more often; but every renewal can bring a product changeover and a setup load. For frequent repetition, the processing time of the parts alone is not enough. The time that remains outside processing and is really usable must be enough for these setups. For this reason the calculation brings the demand load and the setup load together on the same resource and the same time horizon. The result is the initial test of a feasible schedule.

The scope of the setup total must be explicit. How many changeovers are needed to produce three products once depends on the state in which the sequence starts and ends. In a one-time run the final state can be left as it is; in a regularly repeating sequence, the return from the last product to the first is also needed. Adding the start setup and counting the same return once more inflates the load. Not counting the return at all reduces the load. Changeover times can differ by product direction. For this reason, instead of a single general setup time per product, a record showing from which state to which state the changeover goes is used. An unknown changeover is not accepted as zero.

Daily spare capacity is found by subtracting the processing load of the demand from the available time. This spare is not a common and free pool of minutes for all resources. While the machine is idle, the setup worker may be on another job; the time may be split into small pieces, or the required material may arrive late. The calculation first takes the time of the same constrained resource as its basis. Then the interruptibility of the setup, worker availability and physical changeover conditions are examined. Even if the average spare equals the setup load, the schedule may not be feasible. In such a case the formula is not wrong; the time placement constraint, which the formula alone does not cover, has been exposed.

The repeat interval is interpreted as a lower bound. The result can come out as a fractional day; if the business has a rule of renewing the product family only on whole days, the calendar is arranged accordingly. Rounding the mathematical lower bound down exceeds capacity. Rounding up can help in terms of capacity; but it can create larger batches and longer stock waiting. For this reason the shortest interval is not the best decision in every condition. Delivery windows, the coverage time per product and the reliability of the setup improvement are assessed together. The calculation makes the effect of setup reduction visible; it does not permit presenting an unverified time saving as realized capacity.

Formula