Process batches
Separate quantities between setups from transfer quantities, equipment loads and traceability lots; reconcile setup effort, flow time and capacity.
Tags: production, planning, batching, flow
On this page
In the 12-piece example, separate operators set up A and B concurrently in 8 min each. A runs at 1 min/piece, B at 2 min/piece, with no rejects or transfer time. One 12-piece transfer finishes at minute 44. Four 3-piece transfers finish at minute 35, while the process batch remains 12 and each machine still has one setup. The saving comes from overlapping work. B’s processing rate stays the same.
Project Production Institute distinguishes the quantity processed under one setup from the quantity moved downstream at a time. Project Production Institute, Batch
Check material-lot and inspection boundaries separately before splitting work.
Name the batch boundary
| Quantity or identity | Meaning to establish |
|---|---|
| Process batch | Quantity processed between setups at the specified operation |
| Transfer batch | Quantity released or moved to the next operation together |
| Simultaneous equipment load | Compatible items occupying one oven, fixture or other shared cycle |
| Production order | Work authorized or planned under an order identity |
| Traceability lot | Material or production history retained under the applicable identification rules |
An order can span several runs or transfers. Several orders may share a compatible equipment cycle if traceability and release rules allow.
Software terminology can also differ. Infor's Production Scheduling 20.x help defines a process batch as a resource-occupying production activity; it describes parallel batches assigned to individual processes. That is a scheduling object with product transformations and resource claims, not merely a count field matching the general definition above. Translate its documented meaning when exchanging data. Infor, About Process Batches
Setup effort and economic lot sizing
Lean Institute Brasil describes changeover as switching production between products and measures it from the last part of the preceding run to the first good part after the changeover. Define the boundary used locally: tool exchange alone may omit adjustment and first-good-piece work. Lean Institute Brasil, Changeover
For a simple original capacity model with demand D, process-batch size Q and setup duration s, the long-run setup time is (D/Q) × s. This assumes one identical setup per run, uniform quantities and no additional sequence-dependent effects. With 120 units, 12-unit runs and eight-minute setups, ten setups consume 80 minutes. Six-unit runs require twenty setups and 160 minutes. Merely transferring each 12-unit run in groups of three does not create those extra process setups under this model.
Economic production quantity models trade setup cost against holding cost while production and consumption overlap. MIT's finite-replenishment treatment makes that tradeoff explicit. Its calculated lot size does not by itself schedule labor, reserve machine hours or define a transfer quantity. MIT CTL, SC1x Key Concepts, v6.2 Fall 2016, printed p. 47
Original example: smaller transfers, unchanged process batch
Assume an invented 12-piece job passing through machines A and B. Each machine receives one eight-minute setup at time zero, using separate operators, and remains set for the entire job. A processes one piece per minute; B takes two minutes per piece. All pieces are good, transfer time is zero, partial transfers are allowed and B can start as soon as the first transfer arrives. There are no other jobs, breaks or restrictions.
With a 12-piece transfer, A finishes the whole job at minute 8 + 12 = 20. B then takes 24 minutes, finishing at minute 44.
With three-piece transfers, the timeline is:
| Transfer | Available from A | Processing on B |
|---|---|---|
| Pieces 1–3 | Minute 11 | Minutes 11–17 |
| Pieces 4–6 | Minute 14 | Minutes 17–23 |
| Pieces 7–9 | Minute 17 | Minutes 23–29 |
| Pieces 10–12 | Minute 20 | Minutes 29–35 |
The job now finishes at minute 35, nine minutes earlier. Each machine still processes a 12-piece batch with one setup. Combined resource work remains 8 + 12 + 8 + 24 = 52 machine-minutes; elapsed completion time falls because operations overlap, not because that work disappears.
Under the same ideal assumptions, one-piece transfers let B start at minute 9 and finish twelve two-minute pieces at minute 33. This is a comparison of stated timelines, not a guarantee that smaller transfers always improve a real system. Handling trips, transport, inspection holds, limited buffers and shared operators were deliberately excluded.
Szendrovits's original multistage production-quantity paper explicitly models sub-batch transport allowing operations to overlap and includes work-in-process as well as finished-stock holding costs. That broader problem is different from optimizing one operation's setup count alone. Szendrovits, Management Science 22(3), November 1975, pp. 298–308
Respect simultaneous-load and release constraints
Oracle's Release 12.1 Production Scheduling documentation describes batchable machines, ovens and kilns processing compatible items simultaneously. Capacity is represented by occupied spots, with grouping attributes such as temperature and duration. It also notes that some equipment can be opened during processing while other equipment must remain closed until completion. These are explicit model choices, not permission to interrupt a qualified process. Oracle, Creating Production Scheduling Models: Batchable Resources
A simultaneous load therefore needs its own duration and capacity basis. Do not multiply a shared furnace cycle by every piece as though each piece independently occupies the full furnace. Equally, do not divide a cycle by the nominal number of positions unless the actual load, compatibility and release rules support that conversion.
Check fixture capacity, approved load patterns, cleaning, material compatibility, shelf life, inspection release and identification. A physically movable piece may still be unavailable downstream until required acceptance is complete. Preserve the original material identity when transfers split; document any approved relationship when loads combine.
Planning handoff and review
Record the operation, process quantity, permitted transfer quantity, simultaneous-load capacity, setup basis and output-release rule separately. Include routing and part revision, containers, handling resources, calendars, loss assumptions and inspection responsibilities. Distinguish pieces started, pieces transferred and accepted output rather than treating all three as the same quantity.
Compare alternatives on completed good output, elapsed time, setup effort, handling work and inventory. Recheck the whole route when a process or transfer quantity changes. A locally lower setup cost can coexist with longer queues; an earlier first transfer can coexist with no change in bottleneck output rate. The selected operating rule belongs to the organization and process qualification, not to this reference example.
Source versions
Reviewed 2026-11-01. MIT's notes carry v6.2 Fall 2016; the historical multistage paper's publisher dates it November 1, 1975. Oracle's cited collection is Release 12.1, and Infor's link is 20.x documentation; neither is assigned an unverified page issue date. No separate publication date was established for the institute glossary pages. The numerical examples and handoff recommendations are original, bounded planning guidance.