aiwiki.page
English
Economics / Business & Management / material-requirements-planning

Material Requirements Planning

Material requirements planning calculates the quantities and timing of components and materials needed to meet a production schedule.

25 keywords2 linked from13 not yet writtenWritten by AI
InventorySupply ChainStock-keeping un…Data GovernanceTransaction CostMathematical opt…Application Prog…Working CapitalMaterial R…

Material requirements planning (MRP) is a method for translating a schedule of finished-product requirements into time-phased requirements for components, subassemblies, and raw materials. It combines information about product structure, available inventory, outstanding orders, and replenishment times to determine what must be purchased or manufactured, in what quantities, and when orders should be released. Its distinguishing feature is the calculation of component demand from the planned production of the items that use those components, rather than the independent forecasting of every item.

MRP became a foundational technique in computerized manufacturing management because complex products may contain thousands of parts shared across many assemblies. A change in the production schedule can therefore affect numerous purchasing and manufacturing decisions. By coordinating those dependencies, MRP connects customer requirements with factory and supplier activity throughout a supply chain. Its outputs are plans and exception signals, however, not guarantees that materials will arrive or production capacity will be available.

The term originally described a relatively specific materials-planning calculation. It subsequently became part of broader manufacturing and enterprise systems incorporating capacity planning, financial control, purchasing, and shop-floor execution. Understanding MRP requires distinguishing its underlying calculation from these wider systems, and distinguishing a mathematically consistent materials plan from a feasible production schedule.

Historical development and scope

MRP developed from earlier practices for maintaining parts lists, controlling stocks, and scheduling manufacturing orders. Traditional stock-control methods often treated each item separately: when inventory fell to a reorder point, a replenishment order was triggered. That approach could work well for independently demanded goods, but it did not fully exploit the known relationships between finished products and their components. A factory planning to assemble 500 machines could calculate the number of required motors or housings directly rather than infer it from historical withdrawals alone.

The expansion of commercial computing during the 1950s and 1960s made repeated, multilevel calculations practical. Computers could store product structures, maintain large item files, and recalculate requirements across many planning periods. Joseph Orlicky was a central figure in formalizing and explaining this approach. His 1975 book, Material Requirements Planning: The New Way of Life in Production and Inventory Management, became an influential statement of the method. Oliver Wight and George Plossl also contributed substantially to the development and dissemination of manufacturing planning practices.

Early MRP concentrated on material availability. Its limitations encouraged the development of closed-loop MRP, in which capacity checks and feedback from purchasing and production execution were connected to the materials plan. A planned order could then be reconsidered if a supplier could not deliver or a work center could not complete the required operations. “Closed loop” referred to the return of execution information to planning, not to the elimination of uncertainty.

During the late 1970s and 1980s, the broader concept of manufacturing resource planning, usually written MRP II, extended this integration to business planning, financial projections, and other manufacturing resources. The Roman numeral distinguishes it from material requirements planning; the two expressions are not interchangeable. MRP remained a calculation within the larger framework.

From the 1990s, enterprise resource planning (ERP) systems integrated manufacturing with functions such as finance, sales, procurement, and human resources. Many ERP applications retain an MRP engine, but ERP is not simply another name for MRP. An enterprise may use ERP without manufacturing anything, while a manufacturer can perform an MRP calculation in specialized software or even a spreadsheet. The historical progression therefore concerns expanding organizational integration rather than successive replacement of the original arithmetic.

Independent demand, dependent demand, and the production schedule

The conceptual foundation of MRP is the distinction between independent and dependent demand. Independent demand originates outside the product structure being planned, as with customer orders for finished bicycles or replacement pumps. Dependent demand follows from another item's production requirements, as with the wheels needed to assemble those bicycles. The same part may face both kinds of demand: a motor can be used in new equipment and sold separately as a service replacement.

This distinction determines where forecasting belongs. A manufacturer may forecast sales of a finished product, but once its production quantity is specified, the nominal requirement for a component follows from the product structure. If each machine requires four brackets, a production order for 75 machines creates a requirement for 300 brackets before adjustments for scrap, yield, and existing stock. Forecasting bracket withdrawals independently can obscure the timing and quantity already implied by the assembly schedule.

The main interface between market demand and detailed materials planning is the master production schedule (MPS). It specifies quantities and dates for selected finished goods, major assemblies, or other master-scheduled items. It is not necessarily identical to the sales forecast. A forecast describes expected demand; the MPS expresses a production intention after considering orders, inventories, business priorities, and production constraints. In make-to-stock production it commonly schedules finished goods, while in assemble-to-order environments it may schedule modules or options.

Customer orders and forecasts must be reconciled to avoid double-counting. One common mechanism is forecast consumption: actual orders consume an appropriate portion of the forecast within defined periods. If a monthly forecast is 1,000 units and 600 units of qualifying customer orders have arrived, planning should not automatically interpret the two figures as demand for 1,600 units. The exact consumption rules depend on order dates, product groupings, and planning policies.

The schedule also defines the boundary within which MRP operates. A materials calculation can be internally correct but commercially inappropriate if its master schedule contains unrealistic volumes or obsolete priorities. Similarly, a planning horizon that ends before long-lead-time purchases become visible can conceal future shortages. The horizon ordinarily needs to cover the relevant cumulative replenishment time, including linked purchasing and manufacturing stages. Stable near-term commitments, appropriate demand aggregation, and explicit treatment of service demand are therefore central to the quality of the resulting plan.

Product structures and planning records

The principal structural input is the bill of materials (BOM), which records the components and quantities needed to make a parent item. A single-level BOM describes the immediate children of one assembly. A multilevel BOM connects those relationships into a product structure extending from the finished item to purchased materials. If a product contains two modules and each module contains three bearings, the direct usage relationship implies six bearings per product, assuming no other sources of bearing demand.

Real product structures require more than multiplication. Components may be shared by several products, substituted under defined conditions, or introduced through engineering changes. Bills may carry effective dates or serial-number applicability so that the correct revision is used. A planning BOM may represent product families or expected option combinations, whereas a manufacturing BOM describes the structure intended for production. Confusing these purposes can create requirements for combinations that no actual customer ordered.

The inventory record provides the second major body of information. For each stock-keeping unit, and often for each site or planning location, it identifies quantities on hand, open orders, scheduled receipt dates, allocations, and planning parameters. Physical stock is not always usable stock. Material may be quarantined, expired, owned by a customer, assigned to another order, or located where it cannot satisfy the requirement in time. MRP must use the relevant availability rules rather than simply the total quantity recorded across all locations.

Item parameters commonly include procurement type, replenishment lead time, order quantity rules, safety stock, order multiples, and scrap allowances. Units of measure require particular care. A component may be purchased by the kilogram, stored by the roll, and consumed by the meter. Incorrect conversion factors can produce systematic shortages even when every arithmetic operation in the planning engine is correct.

Product structures are normally processed in an order that allows all parent requirements to be collected before a component is planned. Low-level coding assigns an item to its deepest occurrence in the relevant product structures. A shared component is then netted after requirements from all higher-level parents have been accumulated. Without this sequencing, the system could plan a component prematurely and fail to recognize additional demand from another assembly.

These records make MRP strongly dependent on data governance. Engineering changes, warehouse movements, purchase-order confirmations, and production completions alter different parts of the same planning model. If they are updated at incompatible times, a formally consistent calculation may describe a factory that no longer exists in operational reality. Record accuracy is therefore not merely an administrative concern; it is part of the mechanism by which material requirements are calculated.

Explosion, netting, and lead-time offsetting

An MRP calculation combines three operations: explosion, which derives component requirements from parent orders; netting, which deducts usable supply; and offsetting, which places order releases earlier than required receipts by the relevant lead time. These operations are repeated down the product structure. Time is commonly divided into daily or weekly buckets, although some systems retain more precise dates and aggregate them for display.

For an item in period tt, let GtG_t denote gross requirements, StS_t scheduled receipts from existing orders, and Pt−1P_{t-1} projected available inventory carried from the preceding period. Let SStSS_t be a required ending safety-stock level. A basic net requirement is

Nt=max⁡(0,  Gt+SSt−Pt−1−St).N_t=\max\left(0,\;G_t+SS_t-P_{t-1}-S_t\right).

An order-sizing rule converts NtN_t into a planned receipt QtQ_t. The projected balance becomes

Pt=Pt−1+St+Qt−Gt.P_t=P_{t-1}+S_t+Q_t-G_t.

With lot-for-lot sizing, Qt=NtQ_t=N_t. Other rules can make the planned receipt larger. These equations describe a simplified convention; systems differ in how they handle reservations, shortages, safety stock, and receipts within a period.

Consider a product A required in a quantity of 100 in week 6. There are 20 usable units of A on hand and an existing scheduled receipt of 10 units in week 6. With no safety-stock requirement, 70 additional units must be received. If assembling A takes one week, the planned order release is in week 5. Each A requires two units of B and three units of C, both assumed to be needed when assembly starts.

Item Requirement period Gross requirement Opening usable stock Scheduled receipts available Planned receipt Lead time Planned release
A Week 6 100 20 10 70 1 week Week 5
B Week 5 140 30 20 90 2 weeks Week 3
C Week 5 210 10 0 200 1 week Week 4

The scheduled receipt of 20 units of B is assumed to arrive by week 5. Crucially, B and C requirements arise from the 70 units of A newly planned for manufacture, not from all 100 units of finished-product demand. Existing A inventory and the already scheduled A receipt cover the other 30 units. Components for the existing A order should already be represented through that order's material requirements.

In this example, all components are required at the parent's start date. More detailed systems can place a component at the operation where it is consumed. If C is installed late in assembly, its requirement may occur after the initial release of A. This operation-based timing avoids treating every component as necessary at the first production step, but requires accurate routing and operation information.

Lot sizing, buffers, and planning stability

Net requirements do not uniquely determine order quantities. Lot-for-lot ordering supplies exactly the calculated requirement in each period, limiting residual inventory but potentially creating many orders or setups. A fixed-order-quantity rule supplies a standard amount, while an order-multiple rule rounds requirements to a packaging or process increment. Period-order-quantity methods group requirements across a specified number of periods. These choices alter both stock levels and the timing of dependent requirements.

The economic order quantity illustrates the trade-off between ordering cost and inventory holding cost. In its basic stationary-demand model,

Q∗=2DKh,Q^*=\sqrt{\frac{2DK}{h}},

where DD is demand per unit time, KK is the fixed cost of placing or setting up an order, and hh is the holding cost per unit over the same time basis. The expression assumes conditions that often do not hold for time-phased dependent demand. Using it mechanically for a component with intermittent requirements can create stock long before it is needed.

Dynamic lot-sizing methods instead consider requirements by period. The Wagner–Whitin approach, introduced in 1958, determines a minimum-cost ordering plan under a specified finite-horizon model with known demand, setup costs, and holding costs. Heuristics such as Silver–Meal and part-period balancing seek useful compromises with simpler calculations. Even an optimal single-item lot plan does not necessarily optimize an entire factory, because shared capacity, supplier constraints, and interactions among production stages may remain outside the model.

Buffers address uncertainty rather than order economies alone. Safety stock protects against variation in demand, delivery, or yield, while safety time moves a requirement earlier to provide temporal protection. The two are not equivalent. Additional stock can cover a quantity shortfall; additional time can absorb a late delivery only if the material actually becomes available before use. Applying large buffers at every level can also inflate inventories and obscure the underlying sources of unreliability.

A recurring problem is MRP nervousness: relatively small changes in demand or timing can cause large revisions to planned orders. A requirement moving across a period boundary may change an entire lot, which then changes several lower-level component orders. Fixed order multiples and grouping rules can intensify these discontinuities. Repeatedly instructing suppliers to advance, postpone, or cancel orders can create costs that are not visible in the inventory calculation.

Planning time fences limit how readily near-term schedules can change. Inside a frozen interval, changes may require explicit authorization; farther out, the system may permit progressively more flexibility. Firm planned orders can preserve a planner's intended quantity or date across recalculations. These mechanisms trade some responsiveness for stability. Their economic significance lies partly in protecting supplier commitments, setups, and coordination effort—forms of transaction cost that a purely material-balance calculation does not capture.

Capacity, lead times, and production feasibility

Classical MRP answers a conditional question: what material is required if the proposed production schedule is carried out? It does not, by itself, prove that the schedule can be executed. Several planned orders may require the same machine, tooling, or skilled workers at the same time. Materials can therefore be available for a plan that exceeds the factory's productive capability.

A common sequence combines rough-cut capacity checks at the master-schedule level with more detailed capacity requirements planning for individual work centers. Routings specify the operations required, their sequence, and their setup and processing times. If an operation requires a two-hour setup plus 0.1 hour per unit, a batch of 100 units creates 12 hours of nominal work-center load. The calculation must then be compared with available capacity over the relevant periods.

Available capacity depends on calendars, shifts, labor availability, planned maintenance, and assumptions about effective utilization. A machine scheduled for 40 hours does not necessarily provide 40 hours of productive processing. Conversely, some work centers contain parallel machines or flexible staffing. The appropriate capacity representation depends on whether resources are genuinely interchangeable and whether particular products require specific equipment.

Lead time introduces a deeper difficulty. Conventional MRP often offsets orders using a predetermined replenishment time, but actual manufacturing lead time includes waiting, transport, inspection, setup, and processing. Queueing can dominate processing time, especially at heavily loaded resources. Increasing the planned lead time may cause orders to be released earlier; those additional releases can increase work in process and congestion, undermining the intended improvement.

A related relationship is Little's law, which, for a consistently defined stable system, relates average work in process LL, throughput rate λ\lambda, and average flow time WW:

L=λW.L=\lambda W.

This does not supply a scheduling rule, but it clarifies why work in process and lead time cannot be treated as unrelated quantities. Releasing more work without a corresponding increase in throughput generally does not make all orders finish sooner.

Finite-capacity scheduling explicitly respects resource limits when placing operations in time. Such scheduling may be supplied by an advanced planning and scheduling application or another production-control system using mathematical optimization or heuristics. Material dates and capacity dates must then be reconciled iteratively. An overloaded period may lead to overtime, subcontracting, alternative routings, smaller batches, or a revised customer commitment. Each response changes the conditions on which the original materials plan was based.

Planning runs, order control, and execution feedback

MRP normally produces recommendations rather than automatically binding commitments. Typical outputs include planned purchase and manufacturing orders, proposed order releases, and messages to expedite, defer, reduce, or cancel existing supply. A planned order represents a calculated future action. A released purchase order or production order represents an operational commitment with additional commercial or execution consequences.

Two traditional calculation modes are regenerative and net-change planning. A regenerative run recalculates the selected planning population broadly from current inputs. A net-change run recalculates items affected by changes such as new demand, revised inventory balances, or altered product structures. The second approach can reduce processing work, but changes must propagate correctly through dependencies. Modern software may combine broad periodic recalculation with more frequent selective updates.

Pegging identifies the demand associated with a supply or component requirement. A buyer seeing a shortage of 200 bearings can use pegging to determine which assemblies or customer commitments are affected. This is especially important when the same component serves many products. Some systems maintain explicit links between supply and demand; others calculate relationships dynamically. Pegging is a diagnostic connection, not necessarily an immutable allocation of particular physical units.

Execution feedback closes the planning loop. Purchase-order acknowledgments update expected delivery dates; receiving transactions record actual arrivals; production reporting records completions, consumption, and scrap. If 95 usable units emerge from an order intended to supply 100, the missing five may create a new requirement. If a delayed receipt is still shown as arriving on its original date, the next calculation can conceal the shortage rather than reveal it.

Manufacturing execution systems can provide detailed shop-floor status, while warehouse systems report physical movements and storage availability. Interfaces through an application programming interface or other data-exchange mechanism do not by themselves guarantee consistency. Transaction timing, duplicate messages, unit conversions, and responsibility for correcting errors affect the quality of the shared plan.

Exception messages require interpretation. A recommendation to expedite an order by ten days may be physically impossible, and a cancellation recommendation may disregard contractual penalties or material already produced by a supplier. If every minor date movement generates an alert, important shortages can disappear within a large volume of low-value messages. Organizations therefore use tolerances, priorities, and escalation rules to distinguish actionable deviations. The operational purpose of the calculation is not merely to produce more messages, but to connect material consequences with decisions that can still influence outcomes.

Implementation, economic effects, and performance measurement

Implementing MRP is an organizational change as well as a software project. Engineering controls product structures, warehouse personnel maintain stock records, procurement manages supplier commitments, and production management reports work in progress. The system can only reconcile these activities if their definitions and responsibilities are compatible. A technically successful installation may still produce poor plans when departments use different dates, inconsistent item identifiers, or incompatible assumptions about available stock.

Inventory-record verification is particularly important. Cycle counting checks selected items repeatedly rather than relying solely on an annual physical count. Discrepancies can reveal incorrect issues, unrecorded scrap, location errors, or unit-of-measure problems. Counting alone does not correct the process that created the discrepancy; repeated adjustments without investigation may restore a balance temporarily while leaving future planning exposed to the same error.

The economic benefits sought from MRP include fewer material shortages, lower unnecessary stocks, more reliable delivery commitments, and better coordination of purchasing and production. These effects are conditional rather than automatic. Reducing inventory can release working capital, but a reduction that increases shortages may impair output and customer service. Conversely, raising inventory can improve availability while concealing unreliable supply or inappropriate batch sizes.

Costs also move between categories. Smaller lots can reduce average stock but increase purchasing effort, receiving activity, and production setups. Larger lots may exploit economies of scale in ordering or processing while increasing storage, obsolescence, and financing costs. Evaluating MRP policies therefore involves the opportunity cost of capital and capacity, not simply the recorded value of inventory. The cheapest purchase price is not necessarily the lowest cost for the production system as a whole.

Performance measures commonly include inventory accuracy, schedule adherence, on-time delivery, supplier delivery reliability, stockout frequency, expedite activity, and inventory turnover. Each has interpretive limits. A high rate of order completion against frequently revised promise dates may not indicate reliable customer service. High inventory turnover may coexist with shortages of a few critical parts and excessive stocks of others. Aggregate measures need to be examined alongside item-level and order-level effects.

Attributing improvement specifically to MRP is difficult when implementation also changes training, product design, supplier agreements, and transaction discipline. Software adoption is not an isolated treatment. Comparisons must distinguish improvements in the calculation from improvements in the underlying operating process. A useful assessment asks whether plans became more executable, whether exceptions were resolved earlier, and whether gains in productivity or inventory performance were achieved without simply transferring costs to suppliers, customers, or another production stage.

Alternative planning approaches and contemporary extensions

MRP is often contrasted with just-in-time production and pull control, but the distinction is not absolute. MRP usually derives time-phased orders from anticipated requirements, while a pull mechanism authorizes replenishment in response to downstream consumption. Kanban is one means of transmitting such authorization. A factory can use MRP to plan long-lead-time purchases and overall capacity while using pull replenishment for stable, repetitive flows within production.

The approaches fit different operating conditions. A complex, low-volume assembly with long procurement lead times benefits from explicit visibility into future component requirements. A repetitive production line with short, reliable replenishment cycles may control many items effectively through consumption signals and limited container quantities. Product variety, demand variability, setup times, and supplier reliability determine the practical boundary between the methods. Neither the “push” nor “pull” label alone establishes which will perform better.

Demand-driven material requirements planning (DDMRP) is a distinct methodology that places inventory buffers at selected decoupling points and uses buffer-based planning and execution rules. These points interrupt the direct propagation of requirements through the full product structure. DDMRP should not be confused with ordinary MRP that happens to receive updated customer orders; its defining feature is the particular use of strategically positioned buffers and associated replenishment logic. Its performance depends on how those positions and parameters match the operating environment.

Multi-site and globally distributed production extend the basic problem. A component may be purchased centrally, transferred between plants, processed by a subcontractor, and assembled elsewhere. Planning must represent transfer times, location-specific inventories, calendars, and ownership or sourcing rules. Distribution requirements planning applies related time-phased logic to replenishment across a distribution network, although it primarily follows movement between stocking locations rather than component transformation through a manufacturing BOM.

Modern systems also incorporate alternative suppliers, substitute materials, shelf-life constraints, lot traceability, and product configuration. These features make the model more realistic but introduce additional decision rules and potential conflicts. A nominally available substitute may require customer approval; an available batch may expire before use; a supplier with the shortest quoted lead time may lack capacity. Representing such constraints correctly can matter more than increasing the frequency of recalculation.

Machine learning can support the estimation of demand, delivery reliability, or lead times, while simulation can test planning policies under variable conditions. These methods supplement rather than remove the need for explicit material relationships and conservation of quantities. A predicted arrival date remains uncertain, and an accurately forecast component shortage still requires an executable response. Contemporary development therefore concerns both better predictions and tighter coordination among material plans, resource constraints, and actual production events.