How do you calculate net requirements in MRP?

Net requirement equals gross requirement minus on-hand inventory minus scheduled receipts for each period. It is the quantity still needed after existing and incoming stock is used up. The lot-sizing rule then acts on this net-requirement row, deciding whether to order each period exactly, batch several periods, or order a fixed quantity.

What is the Wagner-Whitin algorithm?

Wagner-Whitin is a dynamic-programming method that computes the mathematically optimal set of orders over a finite planning horizon, minimizing total setup plus carrying cost for a single item with known period-by-period demand. It is the benchmark that heuristic rules are measured against, but it is heavier to compute and assumes you trust the forecast across the horizon.

How do you choose the best lot-sizing rule?

Price one setup and one period of holding, read the demand pattern, and weigh the item's value and any hard constraints like supplier minimums or shelf life. Steady demand favors a fixed economic quantity; lumpy demand favors lot-for-lot or period order quantity. Use simple rules for most parts and reserve dynamic or optimal algorithms for high-value items.

Lot Sizing in MRP: Rules and Trade-offs

Q: What is lot sizing in MRP?

Lot sizing is the rule an MRP system uses to turn a stream of net requirements into actual planned orders, deciding how much to order or make each time. It sets how much inventory an item carries and how often it is ordered or set up, and every rule balances the same trade-off: setup and ordering cost against inventory carrying cost.

Q: What are the main lot-sizing rules?

The three you meet first are lot-for-lot, which orders each period's exact net requirement; fixed order quantity, which orders the same preset amount every time; and period order quantity, which batches a fixed number of periods of net requirement. Beyond these are dynamic heuristics such as least unit cost and part-period balancing, and the mathematically optimal Wagner-Whitin algorithm.

Q: What is the difference between static and dynamic lot-sizing rules?

Static rules apply a fixed policy regardless of the demand pattern, like lot-for-lot, fixed order quantity, and economic order quantity. Dynamic rules read the actual pattern of upcoming net requirements and adjust the order timing and size, like period order quantity, least unit cost, and Wagner-Whitin. Dynamic rules can cost less but are heavier to run and depend on trusting the forecast.

Q: Why is EOQ not ideal for MRP?

Economic order quantity assumes steady, continuous demand, but MRP plans dependent-demand components whose requirements arrive in lumps tied to parent-order timing. Forcing a fixed EOQ onto lumpy demand over-orders in thin periods and carries inventory the discrete pattern never needed, which is why discrete rules like lot-for-lot and period order quantity usually fit MRP better.

How MRP turns lumpy net requirements into orders, the menu of rules from lot-for-lot to Wagner-Whitin, and the one trade-off behind all of them.

Lot sizing in MRP is the rule that decides how much to order or make each time a material requirements planning run says you need something. Every rule balances one trade-off: setup and ordering cost against inventory carrying cost. Small lots cut inventory but add setups; large lots cut setups but tie up cash.

MRP calculates what you need and when. Lot sizing answers the question MRP leaves open: now that a net requirement exists, how big should the order be? Pick the rule badly and you either drown in changeovers and tiny orders or bury cash in stock you will not touch for a month. This post explains the net-requirement math lot sizing acts on, walks the main rules from lot-for-lot to the dynamic ones, and shows the single trade-off behind all of them. It is educational and names no products.

What is lot sizing in MRP?

Lot sizing is the policy an MRP system applies to convert a stream of net requirements into actual planned orders. When MRP explodes a bill of materials and time-phases demand, it produces, for each item, a row of net requirements by period: what is still needed after on-hand inventory and scheduled receipts are used up. Net requirement equals gross requirement minus on-hand minus scheduled receipts. Lot sizing decides whether to order each period's net requirement exactly, batch several periods together, or order a fixed economic quantity. The choice sets how much inventory the item carries and how often it is set up or ordered.

The reason lot sizing needs a dedicated rule is that MRP demand is usually lumpy. Because a parent's orders are batched, the components underneath inherit spiky, on-again-off-again requirements rather than a smooth flow. That lumpiness is what makes lot sizing interesting: a rule that is optimal for steady demand can be wasteful for spiky demand. Lot sizing is a core step in production scheduling and turns the output of the master production schedule into buildable, buyable orders.

MRP derives net requirements from gross demand, on-hand, and receipts. The lot-sizing rule turns that lumpy net-requirement row into concrete planned order releases.

What are the main lot-sizing rules?

Lot-sizing rules split into two families: static rules, which use a fixed policy regardless of the demand pattern, and dynamic rules, which look at the actual pattern of upcoming net requirements and adjust. The three you meet first are lot-for-lot, fixed order quantity, and period order quantity.

Lot-for-lot (L4L) orders exactly each period's net requirement, carrying almost no inventory but requiring an order or setup every period there is demand. It is the low-inventory endpoint. Fixed order quantity (FOQ) orders the same preset amount every time, often the economic order quantity or a supplier minimum, carrying the leftover but ordering rarely. It is the low-setup endpoint. Period order quantity (POQ) orders the combined net requirement for a fixed number of periods, then repeats, giving the no-permanent-excess behavior of lot-for-lot with the batching of a fixed quantity. Our companion on lot-for-lot vs fixed order quantity works these three in detail.

Static rules apply a fixed policy; dynamic rules read the upcoming demand pattern. Wagner-Whitin is the mathematically optimal dynamic rule, at the cost of more computation.

Beyond these are the dynamic heuristics. Least unit cost keeps adding future periods to an order while the cost per unit keeps falling. Part-period balancing adds periods until the accumulated carrying cost roughly equals one setup cost. The Wagner-Whitin algorithm computes the mathematically optimal set of orders over the horizon, minimizing total setup plus carrying cost, but it is heavier to run and assumes you trust the forecast that far out. Most plants use a static or simple dynamic rule because it is transparent and robust, and save the optimal algorithm for high-value items where the savings justify the complexity.

Why does EOQ often fit MRP poorly?

Economic order quantity assumes steady, continuous demand, and MRP demand is usually neither. The EOQ formula, Q = the square root of (2 x demand x setup cost divided by holding cost), finds the single batch size that balances ordering and holding cost when demand flows at a constant rate. That assumption holds for independent-demand items pulled smoothly off a shelf. It breaks for the dependent-demand components MRP plans, whose requirements arrive in lumps tied to parent-order timing. Forcing a fixed EOQ onto lumpy demand over-orders in thin periods and can still miss in fat ones, carrying inventory the discrete pattern never needed.

That mismatch is the whole reason discrete rules like lot-for-lot and period order quantity exist. They order to the actual net-requirement pattern instead of to an averaged rate, so they do not stockpile against demand that is not really continuous. EOQ still has a role: it makes a sensible default for the fixed quantity when demand happens to be fairly steady, and it anchors the period count in POQ. But treating it as the universal answer is the classic MRP lot-sizing mistake, and it comes straight from ignoring how lumpy dependent demand really is. Keeping inventory low here is the same instinct that drives lean manufacturing everywhere else in the plant.

How do you pick a lot-sizing rule?

There is no universally best rule, only the rule that fits an item's costs and demand. Work it in this order:

Price the two costs. Put a real number on one setup or order and on holding one unit for one period; their ratio drives everything.
Read the demand pattern. Smooth demand tolerates a fixed economic quantity; lumpy, discrete demand favors lot-for-lot or period order quantity.
Weigh the item's value. Expensive or perishable items push toward lot-for-lot to avoid tying up cash or risking spoilage; cheap, stable items tolerate bigger batches.
Honor hard constraints. Respect supplier minimums, pack multiples, shelf life, and process batch sizes, which can force a fixed quantity regardless of the math.
Match effort to stakes. Use a simple, transparent rule for most parts; reserve dynamic or optimal algorithms for high-value items where the savings pay for the complexity.
Recheck as costs change. Re-decide when setup time drops after a changeover project, when holding cost shifts, or when demand smooths or spikes.

The order rewards the cheap moves. Most items do not deserve a sophisticated algorithm; a clear static rule with honest cost inputs beats a fancy rule fed stale numbers. Save the heavy math for the few parts whose value makes the optimization worth auditing.

What do the standards and data say?

Context from bodies of knowledge and primary data:

Lot sizing and its rules, lot-for-lot, fixed and period order quantity, and dynamic methods, are defined in the supply-chain body of knowledge maintained by the Association for Supply Chain Management (ASCM/APICS) all framed around balancing setup or ordering cost against carrying cost.
The Wagner-Whitin algorithm is the established dynamic-programming method for the single-item dynamic lot-sizing problem, producing the minimum-cost order plan over a finite horizon, and is the benchmark against which heuristic rules are measured.
The base of operations running MRP is large: the Bureau of Labor Statistics reports roughly 13 million manufacturing jobs in the United States, across plants time-phasing thousands of dependent-demand components.

The consistent frame across the standards: lot sizing is a cost-balancing decision, not a formula to memorize, and the right rule depends on the item in front of you.

How does lot sizing ripple through the plan?

Lot sizing is not a quiet back-office setting; it echoes up and down the plan. Large lots on a component create big, infrequent bursts of work at the work centers that make it, which is a direct source of the lumpy load that load leveling then has to smooth. Small lots demand more changeovers, which only pay off if setup time is short. So a lot-sizing policy that ignores capacity can quietly manufacture the peaks and valleys the scheduler fights every week. Good lot sizing and a level, capacity-aware schedule are two halves of the same job.

The same connection runs to pull systems. A part planned with tiny lot-for-lot orders behaves like a pull loop, and sizing its kanban is the shop-floor version of the same trade-off between order frequency and inventory. Whether you express the decision as a lot-sizing rule in MRP or a card count in a kanban loop, you are answering the same question: how small can the batch be before setups cost more than the inventory saves?

Where lot sizing lives or dies: the data underneath

Every lot-sizing rule rests on two numbers plants rarely keep current: the real cost of a setup and the real cost of holding a unit. Setup costs get frozen into routings and never revisited after a changeover project cuts them in half; holding cost gets treated as a fixed corporate percentage while space, capital, and scrap risk drift. Feed a rule stale costs and it recommends the wrong batch with total confidence, almost always a bigger one than the plant now needs, and the excess inventory the rule was meant to control quietly returns. The failure is not the algorithm; it is the inputs. Harmony is an AI-native layer that connects machines, software, and paperwork into one operational layer, with no rip-and-replace, so the numbers behind a lot-sizing decision, actual changeover times, real on-hand and scheduled receipts, current usage, become one live record instead of scattered stale ones. AI search returns cited answers across those records, so a planner can ask what a part's true setup time has been since the last improvement or how much is really on hand and get a grounded answer, and Harmony's digital workflows keep the lot-sizing inputs tied to what the floor is actually doing. It is the same paper-to-digital move Harmony makes elsewhere in the plant (see the CLS case study): the batch size stops being a fossil in a routing and becomes a current, defensible decision.