Workload Control in Manufacturing: When to Release Jobs

Home » Resources » Workload Control in Manufacturing: When to Release Jobs

Estimated reading time: 11 minutes

The cutting machine has room, so another order goes into production. Cutting finishes on time. Then the parts join a growing queue for machining, alongside jobs that should already be moving to assembly.

Starting earlier has made the shop busier without bringing the customer’s delivery closer.

Workload control addresses this problem before another job reaches the floor. Instead of releasing every ready order, you consider the work it will add along its route. That is especially useful in make-to-order manufacturing, where two apparently similar orders can put very different demands on shared resources.

The release decision is only half the job. Accepted work still needs a coordinated sequence through those resources. SkyPlanner APS turns that work into one connected production schedule, reducing the effort of matching a release list to separate machine queues and making the delivery consequences easier to see.

What is workload control in manufacturing?

Workload control (WLC) is a production planning and control approach that regulates work entering production according to the load on its resources. In workload-based order release, orders wait in a pre-shop pool until a release rule permits them to enter the shop floor. Work-centre workload norms help determine which orders can be released.

The purpose is to control work in process (WIP): work already released but not yet completed. A work centre may be one machine or a group of suitably interchangeable resources. Its workload is the work assigned to it, often expressed in estimated processing hours. Its capacity is the ability to perform that work. A workload norm is a limit used by the release policy, not another name for available capacity.

WLC is particularly relevant when products have different routes and processing times: custom machining, fabrication and other high-mix job shops. Counting orders alone tells you little about that mix. One order may require a short visit to cutting and extensive machining; another may bypass machining altogether.

The wider WLC approach can also support enquiry management and due-date decisions. Here, the focus is the daily release question: which waiting orders should enter production now?

Look beyond the queue at the first machine

A machine’s visible queue does not show all the work heading toward it. Released jobs may still be upstream, waiting for or undergoing earlier operations. They will compete for the same downstream capacity later.

For release control, distinguish:

• Direct load: work already at the work centre, waiting or being processed.
• Indirect load: work from released orders that is still upstream but will visit that work centre.
• Pool load: work in orders not yet released. It becomes a candidate addition when you consider an order for release.

Suppose machining looks quiet this morning, but several released fabrication orders will reach it tomorrow. Filling the fabrication area with more machining-bound work ignores that approaching load. Conversely, a waiting order that bypasses machining may be a sensible release candidate.

This is the value of route-based control: it distinguishes jobs that add to a developing queue from jobs that use a different path. “There is space to start” becomes “there is room for this order’s load at the controlled work centres it will visit.”

Define the load calculation before setting the limit

WLC is a family of approaches, not one universal formula. Release methods differ in how they account for downstream work, when they review the pool and how they respond to a work centre running out of work.

One documented method uses corrected aggregate load. Each operation contributes its estimated processing time divided by its position in the order’s original route. The first operation contributes its full time, the second half and the third one-third. That assigned contribution remains in the corresponding work centre’s load until the operation is completed.

The correction distinguishes work that is immediately at a resource from work that will spend part of its released life upstream. It is a load-accounting convention, not a reduction in the processing time the machine must eventually provide.

Do not mix the conventions. A six-hour second operation contributes three corrected load hours in this method, but it still needs six processing hours in the production schedule. A norm calibrated for corrected load cannot simply be reused with a ledger that counts every downstream operation at full time.

Before using a release limit, write down the load unit, which operations contribute, how their contributions are calculated and which completion event removes each contribution. Those definitions make the decision repeatable across planners and shifts.

A workload-based release decision, step by step

Imagine an order waiting for release with this route: cutting for two hours, machining for six hours and assembly for three hours. Using the corrected aggregate convention above, its contributions are two, three and one load hours respectively.

Assume the current load ledger and locally chosen norms are as follows. These numbers illustrate the decision; they are not recommended norm levels.

Controlled work centre	Current corrected load	Candidate contribution	Load if released	Workload norm
Cutting	5	2	7	8
Machining	8	3	11	10
Assembly	4	1	5	6

Cutting and assembly remain within their norms. Machining does not. Under a rule requiring every controlled work centre on the route to remain within its norm, the order stays in the pool. A free cutting machine does not override the machining limit.

At the next review, suppose completed machining work has removed two corrected load hours from that ledger. With no other load changes, machining’s current load is now six; adding the candidate’s three gives nine. The order now passes all three checks.

A practical periodic release procedure follows this logic:

1. Refresh the current load. Remove the assigned contributions of completed operations and reconcile changes to released work.

2. Rank eligible pool orders. Use an agreed rule, such as due date or planned release date, after checking readiness.

3. Test the next order’s route. Add its proposed contributions to the current ledger and compare them with the relevant norms.

4. Commit only an accepted release. Add its contributions to the ledger. If the order fails, keep it in the pool without leaving its proposed load behind.

5. Continue through the candidates. Every later check uses the ledger updated by earlier accepted releases.

That final detail matters. Two orders can each fit against the morning’s starting load but exceed a limit when released together. A rejected order also need not stop all releases: another route may fit, depending on the chosen pool-selection rule.

When an operation finishes, remove its assigned contribution from that work centre. Do not remove the whole order’s load from all later centres simply because its first operation is complete.

Choose a release rhythm that protects flow

Periodic release reviews candidates at set intervals. Continuous release responds to defined load conditions as they occur. Some WLC methods combine the two: a periodic decision balances loads, while a supplementary rule addresses starvation at a work centre.

An empty queue should therefore trigger the agreed policy, not an informal instruction to release whatever is available. A starvation-avoidance rule needs to say which job may enter and how its load is recorded. Otherwise, exceptions can gradually undo the control.

The same discipline applies to urgent orders. Keep pool age and delivery risk visible. If a due order repeatedly fails a norm, decide whether to change capacity, routing, priority or the delivery commitment. Simply relabelling it urgent does not create processing time.

A norm also cannot solve sustained overload. If accepted demand keeps exceeding achievable output at a key resource, holding more orders outside production only changes where they wait. Capacity and customer commitments still need attention.

Tune norms against total lead time, not a tidy shop floor

A very loose norm admits work easily but allows queues to grow. A very tight norm can shorten the time jobs spend on the floor while leaving them waiting too long in the pool. Large jobs can become especially difficult to release.

For an order entering the pool, assess its time to completion as two connected parts:

Pool waiting time + time from release to completion.

Moving a week of waiting from the shop floor into the pool is not automatically a week saved for the customer.

Start by making the load ledger and completion feedback dependable. Introduce or tighten norms gradually, then review whether throughput, delivery performance and total order lead time improve together. Watch the oldest pool orders as well as averages, and investigate work-centre starvation rather than celebrating lower floor WIP alone.

There is no universal “correct” number of load hours. Route mix, capacity, estimation quality and the chosen accounting method all influence the level that works. Revisit it when those conditions change.

WLC, BOA, CONWIP and POLCA: different control questions

These methods all address production flow, but their signals are not interchangeable.

Approach	What the control focuses on
Workload-based WLC release	Whether a candidate’s contributions fit the workload norms of controlled work centres on its route.
BOA	The German belastungsorientierte Auftragsfreigabe tradition: releasing production orders according to capacity loading and a defined loading limit.
CONWIP	Permission governed by a total-WIP limit within a defined production boundary.
POLCA	Capacity authorization through overlapping card loops assigned to pairs of cells, alongside job authorization.

Classic BOA and the different WLC release methods should not be treated as one identical algorithm. In particular, do not import a BOA loading percentage into a corrected-load example without defining how that method measures load and sets its limit.

Choose by the decision you need to control. A total-WIP ceiling limits the amount inside a boundary; work-centre norms distinguish where different routes will put the load. Neither answer, by itself, specifies every operation’s machine and start time.

Turn released work into a connected schedule with SkyPlanner

Passing a workload check means an order can enter under the chosen release policy. It does not mean its next operation can start immediately or that its delivery date is secure. Machines have shifts, materials have availability dates and operations depend on earlier work.

SkyPlanner APS connects those constraints across the accepted work. Its scheduling AI, Arcturus, assigns process steps to eligible workstations using configured capacity, shifts, dependencies and priorities. With material-aware scheduling enabled, material availability is part of the calculation too. The Gantt view shows the resulting plan across resources.

Return to the example: the machining norm now permits the order, but its six-hour operation still needs a place among existing jobs. SkyPlanner calculates that placement together with cutting, assembly and the other orders. You can see the expected completion and the effect of changing priorities without rebuilding each machine queue by hand.

You steer the schedule through priorities, rules and locks; Arcturus does the scheduling calculation. That is the practical reason to use SkyPlanner alongside a workload-based release policy: less manual reconciliation and a clearer view of what the accepted work means for deliveries.

An order’s start eligibility date can carry an agreed earliest-start boundary into scheduling when the eligibility-date rule is enabled. The date is a scheduling restriction, not proof that the order has passed the current WLC norms. Keep that release decision and the load ledger explicit when designing the workflow.

SkyPlanner’s Job Report and Order Item Report also help you review recorded work and duration differences. They give you a basis for improving the time estimates behind planning; the release policy still needs its defined completion feedback and accounting rules.

What better timing changed at Mamec

Mamec, a manufacturer of industrial mixing equipment and high-tech contract products, reports a clearly lighter production-planning workload after adopting SkyPlanner. Its customer account also describes more sensible lead times and less working capital tied up unnecessarily in WIP because jobs start at the right time.

That is the scheduling outcome worth pursuing alongside release control: work beginning when it can progress, rather than simply joining another queue. For a planner working with release norms, the same distinction matters: a controlled start should lead into a coordinated plan, not another round of queue repair.

Start with one release decision and follow it to delivery

For a first implementation, choose a meaningful set of controlled work centres and representative order routes. Confirm that grouped machines really can share the work, establish the completion feedback and agree how to handle a job that cannot pass its norms in time. Then test whether your release decisions and production schedule tell a consistent story.

SkyPlanner integrates with any ERP or MES through its open REST API, so improving that schedule does not require replacing the system that holds your orders.

To see the connection in your own production, book a SkyPlanner demo. Bring a waiting order’s route, the current work-centre loads and your intended release rule. Follow the accepted work through the schedule and examine its effect on completion dates and competing orders.

Prefer to explore first? Try SkyPlanner with demo data or use the AI-assisted Getting Started flow in SkyPlanner Assistant. See how much easier planning becomes when the full route is calculated in one place, while you stay in control of the priorities.