Material requirements planning (MRP), explained with a working calculator

What MRP is

MRP exists because demand for components is dependent: nobody orders 400 springs for their own sake. They are needed because 200 arm assemblies must be built, because lamps are due. Forecasting each component separately ignores that link. MRP calculates component demand directly from the end-item plan instead. Joseph Orlicky and colleagues developed the approach at IBM in the 1960s, when computers first made the arithmetic practical for thousands of items.

The three inputs

Every MRP run depends on three sets of data. If any one is wrong, the plan is wrong.

1. Master production schedule (MPS) — how many of each end item must be completed in each period. It is the only independent demand MRP sees; everything below it is derived.
2. Bill of materials (BOM) — the parent–component structure and the quantity of each component per parent (“quantity per”). In the example, 1 lamp needs 1 arm assembly and 1 arm assembly needs 2 springs, so 1 lamp needs 2 springs.
3. Inventory records — for every item: on-hand stock, scheduled receipts (orders already released and not yet received), lead time, safety stock and the lot-sizing rule.

In practice, data accuracy decides whether MRP is trusted. A BOM missing one component or an inventory record that is 10% off produces plans that planners quietly ignore, which is why cycle counting and engineering change control sit alongside every MRP implementation.

The MRP record, line by line

MRP keeps one time-phased record per item. Each column is a planning period (here, a week). The six lines are calculated left to right:

For each period t, the gross-to-net logic is:

available(t)        = projected on hand(t−1) + scheduled receipts(t)
net requirement(t)   = max(0, safety stock − (available(t) − gross(t)))
planned receipt(t)   = lot size(net requirement(t))
projected on hand(t) = available(t) − gross(t) + planned receipt(t)
planned release(t − lead time) = planned receipt(t)

Worked example: following one order down the BOM

These numbers come from the baseline scenario in the calculator above.

Level 0 — the lamp

The lamp starts with 20 in stock. In week 2, 30 lamps are due but only 20 are available, so the net requirement is 10. Under lot-for-lot, MRP plans a receipt of exactly that quantity in week 2. Final assembly takes 1 week, so the order is released in week 1.

Level 1 — the base, with a fixed order quantity

That lamp release becomes a gross requirement for 10 bases in week 1: the parts are needed when assembly starts, not when it finishes. This is the most common mistake in hand-worked MRP. The base has stock to cover it. The first shortfall arrives in week 3: a net requirement of 20. The supplier sells in lots of 50, so MRP plans a receipt of 50, leaving 30 in stock to cover later weeks, and releases the order 2 weeks earlier, in week 1.

Level 1 — the shade, with safety stock

The shade carries a safety stock of 5. In week 3 the projected balance would fall below it, so the net requirement includes the top-up: 30, not just the raw shortfall. Projected on hand then never drops below 5.

Level 2 — the spring, with a scheduled receipt and a period order quantity

Springs are needed at 2 per arm assembly, so the arm's releases are doubled into spring gross requirements. An open order of 60 springs is already due in week 2, and MRP counts it before planning anything new. The spring uses a two-week period order quantity: when a shortfall appears, one order covers that week and the next. In week 5 that order is 130 springs, fewer orders and more stock than lot-for-lot.

Now switch the calculator to “Shade supplier lead time slips to 4 weeks”. The demand has not changed, but shade orders that should already have been placed show up as past due. That exception is the planner's cue to expedite, find another supplier or tell the master scheduler that the plan is no longer feasible.

Lot-sizing rules

Once MRP knows the net requirement, a lot-sizing rule decides how much to actually order. The choice trades ordering or setup cost against holding cost.

Lot sizes at higher levels amplify demand at lower levels: a fixed lot of 50 bases, or an arm assembly batched for two weeks, turns smooth lamp demand into lumpy component demand. Change the arm's rule in the calculator and watch the spring record move.

Lead-time offsetting, low-level codes and past-due orders

Lead-time offsetting moves each planned order receipt earlier by the item's lead time to produce its release date. Lead times stack: in the baseline, a lamp due in week 8 is started in week 7, so its arm must be started in week 6 and that arm's springs ordered in week 5. The cumulative lead time of the longest path through the BOM sets the shortest horizon on which the plan can respond to change.

Low-level codes make MRP plan each item once. An item's low-level code is the deepest level at which it appears in any bill of materials. Processing items in low-level-code order guarantees that all parents of a component, including a part used both in a sub-assembly and directly in the end item, have been planned before the component is netted.

A past-due planned order is one whose release date has already passed. MRP cannot plan it in time. It reports the exception and still passes the parent's component needs down as immediate requirements, because the parts will be needed the moment the order is worked.

What MRP produces

• Planned orders — purchase requisitions for bought items and work orders for made items, with release and due dates.
• Release messages — planned orders whose release date is now, ready to be converted into real orders.
• Reschedule messages — reschedule in when an open order is needed earlier, reschedule out or cancel when it is no longer needed on its date.
• Exceptions — past-due releases, lead times that cannot be met, negative projected stock.

A planner's job is not to read every record but to act on these messages, and to recognise when the inputs, not the orders, are what need fixing.

MRP vs MRP II vs ERP

When people search for an “MRP system” today they usually mean software that runs this calculation, either a stand-alone MRP product for a small manufacturer or the planning module of an ERP. The arithmetic is the same either way.

Limitations every planner should know

• Infinite capacity. MRP plans as if every work centre can do whatever is asked. A feasible material plan can be an impossible capacity plan, which is why it is checked with capacity requirements planning.
• Fixed lead times. Real lead times vary with load and supplier performance; planners pad them, which inflates inventory.
• Nervousness. Small changes to the MPS can reshuffle many planned orders at lower levels. Firm planned orders and time fences damp this.
• Garbage in, garbage out. Every error in the BOM or the stock record propagates through every level below it.

These limits are behind later approaches such as demand-driven MRP (DDMRP), which places strategic stock buffers to decouple levels of the BOM.

Teaching MRP: exercises to set with this calculator

MRP is usually taught as a hand calculation on a static table. That builds the mechanics but not the judgement. These exercises use the calculator above to get at the judgement:

1. Predict, then check. Before changing week 4 lamp demand from 40 to 70, predict the new spring releases. Explain any difference.
2. Find the minimum lead time slip that breaks the plan. Increase each purchased item's lead time one week at a time. Which item fails first, and why that one?
3. Lot sizing trade-off. Compare L4L, FOQ 50 and POQ 2 for the base. Count orders and sum projected on hand across the horizon. Which would you choose at a $40 order cost and $0.50 per unit-week holding cost?
4. Safety stock where it matters. With the pulled-in order scenario, can safety stock on one item remove all past-due exceptions? Which item, and how much?
5. Write the message. In the lead-time scenario, write the email you would send the master scheduler, with the numbers that justify it.

Frequently asked questions

References and further reading

• Orlicky, J. (1975). Material Requirements Planning: The New Way of Life in Production and Inventory Management. McGraw-Hill.
• Ptak, C., & Smith, C. (2011). Orlicky's Material Requirements Planning (3rd ed.). McGraw-Hill.
• Jacobs, F. R., Berry, W. L., Whybark, D. C., & Vollmann, T. E. (2018). Manufacturing Planning and Control for Supply Chain Management: The CPIM Reference (2nd ed.). McGraw-Hill.

The calculator implements the standard gross-to-net record described in these texts. Its source and tests are deterministic: the same inputs always produce the same plan.