Constraint Planning
Test whether a supply plan is feasible against supplier, labor, machine, storage and transport limits, then compare recovery scenarios without hiding trade-offs.
Each block is planned for 60 minutes: about 35 minutes to learn, 15 minutes for the worked case and 10 minutes for notes or review. Take breaks between blocks. You can complete the lesson over several days.
The idea in simple words
A plan can be mathematically balanced and still impossible to execute. Constraint planning checks the quantities and dates against finite resources. It identifies what binds first, protects decisions from impossible promises, and helps teams compare feasible alternatives.
Start here: Use the time-phased make, buy and transfer proposals from supply planning. They need resource requirements, calendars and dates before load can be compared with capacity.
1. Define feasibility and the constraint scope
Feasibility means the plan respects the limits that matter at the required detail and time. A monthly plant total may conceal a week with too much load; an aggregate line may hide a machine-specific skill or tool requirement. Define the planning horizon, resource level, time bucket, priority rules and whether capacity is infinite or finite. Identify which constraints are hard and which can be exceeded with an approved action such as overtime.
Worked example
The plan needs 80 units in week 2. Monthly capacity is 400, which appears sufficient, but the week-2 line has only 76 standard slots. A monthly view alone would miss four units of overload.
Try it · 10 minutes
Choose a facility and list its capacity resources, time bucket, hard limits, soft limits and the owner who can authorize exceptions.
Watch for: A capacity number is meaningful only with a resource, unit, calendar and period.
2. Build a trustworthy capacity calendar
Available capacity begins with a calendar: shifts, operating days, holidays, maintenance and planned downtime. Convert it to consistent units such as machine-minutes, labor-hours or supplier units per week. Adjust nominal capacity for known efficiency or yield only when the factor is measured and its meaning is clear. Distinguish rated capacity from demonstrated output and from capacity actually available to the plan. Do not count the same crew or machine twice across competing plans.
Worked example
One line has two 8-hour shifts across five days: 80 scheduled hours. A 6-hour maintenance stop leaves 74 gross operating hours. If a validated planning factor is 90%, the indicative effective capacity is 66.6 hours; the factor should not be used to promise a precise output without validation.
Try it · 10 minutes
Create a weekly calendar from shifts and maintenance. Show how each subtraction changes the final available hours.
Watch for: Calendar errors often look like planning errors because every load calculation uses the calendar.
3. Convert each order into resource load
Resource load depends on setup time, run time, quantity, routing, batch rules and yield. A simple equation is load = setup time + quantity × run time per unit. If a product uses multiple operations or resources, calculate each operation’s load in its own units. Use the required quantity at the operation, not automatically the finished-good quantity; yield loss may increase upstream requirements. Keep resource alternatives and their qualifications explicit.
Worked example
A run of 80 bottles takes 3 minutes each and a setup takes 90 minutes. Total line load is 90 + 80×3 = 330 minutes, or 5.5 hours. If capacity is 300 minutes for the week, the plan is 30 minutes over before other products are counted.
Try it · 10 minutes
Calculate setup plus run load for two orders competing for one machine. Compare the sum to the resource calendar.
Watch for: A unit of product is not a unit of capacity; use the correct rate and operation.
4. Identify bottlenecks and dependency effects
A bottleneck is the resource that currently limits throughput for the plan being considered. It can move when mix, dates or supply change. The Theory of Constraints approach often protects the bottleneck’s productive time and uses a buffer against uncertainty; it does not mean every non-bottleneck should run at full utilization. Drum-buffer-rope is one way to coordinate a system around its constraint: the drum sets pace, a buffer protects it and a release rope avoids flooding the system. Apply these ideas carefully to the actual process.
Worked example
If the filler can produce 100 units but the cap supplier can support only 72 kits, the supplier is the current limit. If alternate caps arrive, a labeler with 80 slots may become the next constraint. Expanding the filler alone would not raise output today.
Try it · 10 minutes
Map three sequential steps. Mark each capacity and determine the maximum flow. Then increase the first step’s capacity and see whether the bottleneck moves.
Watch for: Do not optimize a local machine in isolation when another resource controls total flow.
5. Include external and shared limits
Suppliers, docks, labor skills, warehouse slots, packaging equipment, carrier pickups and transport lanes can constrain the same plan. Capacity may be shared across products, customers or plants. Model the resource at the narrowest level that changes the decision, then roll up for reporting. Supplier allocation is not the same as total market capacity; a promised quantity may still depend on approval, raw materials and ship date. Check cross-resource dependencies and avoid counting a shared resource twice.
Worked example
A supplier can send 72 kits per week and a dock can receive 60 pallets, while the line can assemble 76 units. Depending on packaging and conversion, the receiving dock may bind before the production line. A plan that uses units for all three resources without a conversion rule is unsafe.
Try it · 10 minutes
Draw an end-to-end map of one product’s resource limits. Add the unit for each limit and explain how units are converted.
Watch for: External confirmation and internal capacity must align on the same period and quantity basis.
6. Prioritize scarce capacity transparently
When demand exceeds capacity, choices may include priority by customer promise, margin, criticality, fairness rule or contractual obligation. Capacity allocation is a business decision, not merely a solver setting. Make the objective and constraints visible, test the consequences for other products, and document overrides. A feasible allocation may still be undesirable if it creates excess changeovers, inventory or service risk elsewhere. Consider alternate routes, overtime, subcontracting, split lots or date changes only when they are qualified.
Worked example
The line has room for 76 units, but 80 are due. If four units are delayed, the planner needs an approved rule to decide which orders move. “First in spreadsheet” is not a defensible rule unless the business intentionally chose it.
Try it · 10 minutes
Set a transparent allocation rule for a 4-unit shortage and run it against three orders with different due dates and customer classes.
Watch for: The technical plan should expose whose demand is displaced and why.
7. Compare recovery scenarios and protect stability
A recovery scenario changes a defined input: overtime, alternate source, route, shift, priority or due date. Compare each option on feasible quantity, cost, timing, service and risks. Keep the baseline so decision-makers can see what changed. Respect a frozen or firm horizon where changes destabilize work already released. Run sensitivity checks when the input is uncertain. A scenario is not a commitment until the owner approves it and the required parties confirm it.
Worked example
The base plan has a 72-kit supply limit and 76 line slots. An approved alternate source adds 8 kits but takes 5 days longer; overtime adds 4 line slots at a premium. Each may recover a different part of the 8-unit gap. Show the exact dates and whether either option is usable before shipment.
Try it · 10 minutes
Build a three-column comparison: base, alternate supply, overtime. Show the quantity recovered, usable date, cost assumption and remaining gap.
Watch for: Avoid a “what-if” result that forgets approval, lead time or downstream constraints.
8. Monitor capacity health and close the loop
Track overload by resource and period, utilization, queue or backlog, schedule changes, lost production time and the age of unconfirmed supply. High utilization is not always good: a fully loaded system has little room to recover from variability. Compare planned load with actual output and investigate repeated gaps in calendars, rates, setup assumptions or yield. Feed confirmed execution changes into the next plan while keeping a clear audit trail.
Worked example
If a work center repeatedly plans 300 hours but produces only 240, investigate downtime, mix, labor availability and rate assumptions before adding more capacity. A 100% plan utilization with frequent disruptions may produce less throughput than a buffered plan.
Try it · 10 minutes
Write a weekly constraint review with the top binding resources, affected orders, approved recovery actions, owner and next update.
Watch for: Capacity assumptions should be calibrated from evidence, not adjusted until the plan appears feasible.
Further reading and scope
These sources provide background for the planning concepts. The explanations, diagrams and Northstar examples on this page are written for this site. Official software manuals describe the named product and configuration; they do not define a universal process. No source text or figures are reproduced here.
- SAP Help Portal, “Capacity Leveling in MPS and MRP.” Official vendor documentation; publication date not displayed; accessed 30 October 2026.
Describes capacity requirements and capacity planning relationships in a named SAP planning context. - Oracle, “Overview of Material Requirements Planning.” Official product documentation; publication date not displayed; accessed 30 October 2026.
This Oracle guide notes the capacity assumptions used by that MRP product. Capacity behavior varies by planning configuration.
Quick knowledge check
Choose an answer before opening the explanation. These questions review the key ideas from all eight blocks.
1. What makes a capacity value useful?
- A large number
- A named resource, unit, calendar and period
- A color label
- A yearly revenue figure
Reveal answer and explanation
Answer: B. A named resource, unit, calendar and period Capacity must share a meaningful basis with the load being tested.
2. A setup is 90 minutes and 80 units need 3 minutes each. What is total load?
- 170 minutes
- 240 minutes
- 330 minutes
- 390 minutes
Reveal answer and explanation
Answer: C. 330 minutes 90 + 80×3 = 330 minutes.
3. Why can monthly capacity hide a problem?
- Months contain no weeks
- A specific week may be overloaded even when monthly total is below capacity
- Capacity cannot be measured monthly
- Demand always shifts
Reveal answer and explanation
Answer: B. A specific week may be overloaded even when monthly total is below capacity Timing and buckets can hide short-period overloads.
4. What is a bottleneck?
- The most expensive finished product
- The current resource that limits the flow under the plan
- A forecast formula
- Any idle machine
Reveal answer and explanation
Answer: B. The current resource that limits the flow under the plan The binding constraint controls throughput for the scenario.
5. In drum-buffer-rope, what does the drum represent?
- The pace set by the system constraint
- Supplier price
- Finished-goods inventory
- The customer forecast
Reveal answer and explanation
Answer: A. The pace set by the system constraint The drum sets the system pace around its constraint.
6. Why must shared labor not be counted twice?
- It changes the SKU
- Two plans could promise the same finite hours
- It increases forecast seasonality
- It removes setup time
Reveal answer and explanation
Answer: B. Two plans could promise the same finite hours A shared resource is available only once across competing work.
7. What should an allocation rule make visible?
- Only machine utilization
- Whose demand is displaced and the reason
- Only total forecast
- The logo color
Reveal answer and explanation
Answer: B. Whose demand is displaced and the reason Scarce capacity decisions have service and fairness consequences.
8. What should a recovery scenario include?
- Only a new quantity
- Feasibility, dates, costs, risk and remaining gap
- A hidden assumption
- A changed page title
Reveal answer and explanation
Answer: B. Feasibility, dates, costs, risk and remaining gap An option must be usable in time and must show its trade-offs.
9. Why may high utilization be risky?
- It guarantees no downtime
- It leaves little room to absorb variability or disruption
- It reduces demand
- It changes inventory units
Reveal answer and explanation
Answer: B. It leaves little room to absorb variability or disruption A fully loaded resource may have no capacity to recover from normal variation.
10. What should happen if planned hours repeatedly exceed actual output?
- Raise forecast until it looks balanced
- Investigate calendars, downtime, mix, rate and yield assumptions
- Ignore actual output
- Delete the resource
Reveal answer and explanation
Answer: B. Investigate calendars, downtime, mix, rate and yield assumptions Repeated gaps point to assumptions or execution loss that should be diagnosed.
Educational content only. All quantities, products, scenarios and costs are illustrative. Real methods depend on data quality, local policy, system design, contracts and operating constraints.