What Is a Master Production Schedule (MPS)? Formulas and Examples

A master production schedule (MPS) is the committed, time-phased build plan for finished goods: how many of each end item the plant will make in each week. It sits between demand (customer orders and the forecast) and everything downstream. MRP explodes it into component orders, and the shop schedule turns it into jobs on machines. A planner owns it and commits it. That is the difference between an MPS and a forecast: the forecast is a guess about demand, and the MPS is a decision about supply.

Written by User Solutions, Inc., makers of EDGEBIC (RMDB 2.0) by User Solutions. We have built planning and scheduling software for manufacturers in Michigan and around the world since 1991. Reviewed by the User Solutions planning team. Updated September 2026.

This page is the complete reference: the definition, where the MPS sits in the planning chain, the two formulas every master scheduler uses, time fences, and three worked examples you can check line by line (make-to-stock, make-to-order and available-to-promise). If you want to build one yourself, the free master production schedule Excel template does the same arithmetic, and the step-by-step guide to building an MPS covers the weekly process.

The MPS in one picture

Planning runs from broad to detailed. The MPS is the first level where the plan names real products and real weeks.

The MPS takes demand in and passes a committed plan out:

• Inputs: customer orders, the forecast, on-hand inventory, open build orders (scheduled receipts), safety stock, lot sizes, and time fences.
• Outputs: the MPS quantity per item per week, the projected available balance, the available-to-promise quantity for sales, and the demand that MRP explodes.

If you only remember one rule, remember this: the MPS is stated in end items the customer buys (or the option modules they are built from), never in raw materials. Components are MRP's job. For how MRP takes the MPS from there, read what is MRP.

Why it has to be committed by a person

A planning system can calculate a suggested MPS in seconds. The value of the MPS comes from a planner looking at that suggestion and deciding. Three things only a person knows well:

1. Which demand is real. A big quote that sales rates at 50 percent should not drive a full production lot inside the lead time.
2. What the plant can absorb. A suggestion that needs the paint line at 130 percent next week is a wish, not a plan.
3. What the business wants to hold. Building ahead for a summer shutdown or a price increase is a policy choice, not arithmetic.

So the MPS is usually shown in stages. A suggested quantity is calculated but not yet accepted. A firm (committed) quantity is accepted by the planner, and automatic runs will not change it. A released quantity has become a real work order on the schedule. The stages let the far weeks stay soft while the near weeks are locked.

The two formulas every master scheduler uses

1. Gross requirements with a demand time fence

Inside the demand time fence (the next one to three weeks), only customer orders count. Forecast is ignored because the orders you actually hold are the truth that close in, and counting forecast as well would double the same demand. After the fence, the larger of forecast and customer orders counts.

Gross requirement (inside the demand time fence) = customer orders

Gross requirement (after the fence) = the greater of forecast and customer orders

2. Projected available balance (PAB)

The PAB is the stock you expect at the end of each week if the plan runs as written.

PAB (week t) = PAB (week t-1) + MPS quantity due in week t - gross requirement (week t)

Week 1 starts from on-hand inventory. When the PAB would fall below safety stock, the MPS places a lot in that week. With a fixed lot size, it places one lot (or as many lots as it takes) to bring the balance back to at least safety stock.

A third formula, available to promise, gets its own example below.

Worked example 1: make-to-stock

A valve maker builds the HV-40 hydraulic valve to stock. The planning parameters:

• On hand at the start of week 1: 260
• Safety stock: 50
• Fixed lot size: 200
• Demand time fence: 2 weeks (weeks 1 and 2 count customer orders only)

Step 1 is to work out the gross requirement for each week using the fence rule.

Weeks 1 and 2 are inside the fence, so they take the orders (120 and 90), even though week 2's forecast was higher. From week 3 on, the forecast is larger than the orders booked so far, so the forecast counts.

Step 2 is to roll the projected available balance forward, week by week, and add a lot of 200 whenever the balance would drop below 50.

The finished MPS row reads:

Check the arithmetic. Starting stock 260 plus four lots of 200 (800) is 1,060 units of supply. Total gross requirements are 120 + 90 + 110 + 110 + 120 + 120 + 120 + 120 = 910. 1,060 - 910 = 150, which matches the week 8 PAB.

What the planner does with it. The calculation is only a suggestion. Two things jump out:

• Weeks 3 and 4 both need a lot. That is 400 valves in two weeks. If the HV-40 takes 0.2 hours of assembly per valve, each lot is 40 hours. If the assembly cell has only 40 hours a week for this family, two lots in a row is fine, but if the cell also runs other products that week, it is not. This is the rough-cut capacity check: load hours = MPS quantity x hours per unit, compared with the hours the key resource actually has. See rough-cut capacity for the method.
• Week 3 is only one week past the fence. If the HV-40 takes two weeks to build, the week 3 lot must already be in progress. A lot that the calculation "needs" inside the lead time is a warning to act today, not a plan.

The planner firms the week 3 and week 4 lots, leaves weeks 6 and 8 as suggestions, and moves on.

Worked example 2: make-to-order

A conveyor builder makes every unit to a customer order, so there is no forecast to net and no finished stock. The MPS here answers a different question: which week can each order really be built in?

Final assembly is the constraint. It has 48 hours a week, and each conveyor needs 8 hours, so the cell can finish 6 conveyors a week (48 / 8 = 6).

The order book, by the week each customer asked for:

Requested load by week: week 1 has 7 units (56 hours), week 2 has 7 units (56 hours), week 3 has 3 units (24 hours) and week 4 has 4 units (32 hours). Over four weeks the cell has 4 x 48 = 192 hours against 168 needed, so the month is feasible. Weeks 1 and 2 are not: each is one conveyor over.

Loading orders in requested-week order, 6 units per week at most, gives the MPS:

What it tells the planner before anything is late:

• SO-502 and SO-504 each finish one week later than requested, unless something changes.
• The fix costs one conveyor of extra capacity in week 1 and one in week 2: 8 hours of overtime each, 16 hours in total. With it, week 1 builds all 7 requested units and week 2 builds its 7.
• If overtime is not available, the planner calls the two customers now with a real date, instead of discovering the slip on the shipping dock.

Beyond the order book, many make-to-order shops add a small forecast line in the far weeks, for example "5 conveyors a week from week 6", so long-lead parts such as gearmotors are bought before the orders arrive. As real orders come in, they consume that placeholder. This is the same greater-of rule from example 1, used for materials instead of finished stock.

Worked example 3: available to promise

Go back to the HV-40 valve. Sales has a new customer who wants 150 valves in week 2. Can they have them?

The projected available balance cannot answer that, because it mixes forecast (which is not a real customer) with orders. Available to promise (ATP) uses only the supply you have planned and the customer orders you have already accepted.

Discrete ATP is calculated only in weeks that receive supply:

ATP (week 1) = on hand + MPS due in week 1 - customer orders due before the next MPS week

ATP (later MPS week) = MPS quantity - customer orders due from that week until the next MPS week

Using the example 1 plan (on hand 260, lots in weeks 3, 4, 6 and 8) and the customer orders already booked:

How each number is built:

• Week 1: 260 on hand + 0 - orders in weeks 1 and 2 (120 + 90 = 210) = 50
• Week 3: 200 - week 3 orders (70) = 130
• Week 4: 200 - orders in weeks 4 and 5 (40 + 30 = 70) = 130
• Week 6: 200 - orders in weeks 6 and 7 (10 + 0) = 190
• Week 8: 200 - 0 = 200

Check it. Total planned supply is 260 + 800 = 1,060. Total booked orders are 120 + 90 + 70 + 40 + 30 + 10 = 360. 1,060 - 360 = 700, which matches the last cumulative ATP.

The answer to sales. Cumulative ATP through week 2 is only 50, so 150 in week 2 is not possible without taking stock from customers who already have a promise. There are two honest offers:

1. 50 in week 2 and 100 in week 3 (week 3 has 130 available).
2. All 150 in week 3 (cumulative ATP through week 3 is 180).

If the customer takes option 2, cumulative ATP through week 3 drops from 180 to 30. In discrete terms, week 3's 130 is used up and the remaining 20 is taken from week 1's 50, leaving 30 in week 1 and 0 in week 3. Note that ATP ignores safety stock: safety stock protects against forecast error, and promising it away is a separate business decision.

The available to promise glossary entry and available to promise in EDGEBIC cover the discrete and cumulative columns in more depth.

Time fences: how the MPS stays stable

Without rules about when the plan may change, an MPS is rewritten every time a salesperson calls, MRP reissues purchase orders every night, and the floor stops believing the schedule. Time fences fix that.

Two fences define those zones:

• Demand time fence. Inside it, forecast is ignored and only orders count. It is usually set to the final assembly lead time.
• Planning time fence. Inside it, the system will not change planner-committed quantities by itself; it can only send a message suggesting a change. It is usually set to the cumulative lead time: the time to buy the slowest material and build through every level of the bill of materials.

Set the fences per product, not once for the plant. A product with a 10-week imported casting needs a much longer planning fence than one built from stock parts in two days. For a longer treatment with MRP examples, read MRP time fences and planning zones, and for how EDGEBIC by User Solutions applies both fences per product, read firm demand, forecast and time fences in EDGEBIC.

MPS vs MRP vs production schedule

These three are often confused because they all produce "a schedule". They answer different questions.

The gap people fall into is the last column. An MPS that passes a rough-cut check can still be impossible at the bottleneck in a given week, because rough-cut works in weekly totals and ignores sequence and setups. That is why the MPS feeds a finite capacity schedule, not straight to the floor. We cover that handoff in MRP vs APS.

How to build your first MPS in six steps

1. List the end items you will plan. Start with the 20 percent of products that carry most of the volume.
2. Set parameters per item: on hand, safety stock, lot size, demand time fence and planning time fence.
3. Load demand: customer orders by due week and the forecast by week.
4. Calculate gross requirements, PAB and the suggested MPS with the two formulas above.
5. Check it against capacity at the one or two resources that limit output, then adjust.
6. Commit the near weeks and publish ATP to sales.

The master production schedule Excel template has all six steps built in with a worked pump example, and the MPS guide explains how to run the weekly review that keeps it honest.

Common MPS mistakes

• Overstating the plan. Loading more than the plant can build makes every week "behind" from the start. The plan should be believable.
• Planning components in the MPS. The MPS is for end items. Put components on it and you double-count what MRP already calculates.
• No time fences. Every change lands in the next few days and purchase orders churn.
• Forecast inside the demand time fence. It doubles demand that is already on the order book.
• Treating the calculated MPS as the plan. The suggestion is where a planner starts, not where the plan ends.
• Never measuring it. Track how many committed MPS lines were built in the week they were planned. If adherence in the frozen weeks is below about 90 percent, the plan is not realistic or execution needs work.

How EDGEBIC handles the MPS

In EDGEBIC (RMDB 2.0) by User Solutions, the master production schedule is a working grid. Each row is a finished item, each column is a bucket, and each cell shows gross requirements, scheduled receipts, projected balance and the committed quantity. Gross requirements use the greater-of rule with a per-product demand time fence, and a per-product planning time fence protects committed quantities from automatic revision.

Entries move through the three stages described above. Firming records the planner's decision without creating work. Releasing a firm entry creates a build-to-stock work order and links the two, so the master schedule and the floor stay connected. From there the order goes to the finite capacity engine, the same scheduling engine behind RMDB's 35 years of customer results, and is placed on real work centers inside their real shift hours. Available to promise is computed from the same records, so sales sees what the plan can actually support.

Because EDGEBIC works beside your ERP instead of replacing it, the orders, forecast and items behind the MPS come in through Excel, CSV and database import masks. EDGEBIC APS is a one-time $25,000 license, and EDGEBIC Complete, which adds MRP, inventory, purchasing and pegging, is $35,000 one-time. See master production schedule software for the product view, pricing for the editions, and the EDGEBIC MPS explained for the screen-by-screen tour. If you run RMDB today, RMDB remains supported and EDGEBIC is its continuation.

Related reading

• Master production schedule bucket: the period structure an MPS is built on
• MRP formula: gross-to-net requirements: what happens to the MPS next
• What is capacity requirements planning: the detailed capacity check after MRP
• What is advanced planning and scheduling: the system that turns a committed plan into a runnable schedule
• Manufacturing glossary: every planning term in one place