Both models minimize total inventory cost — but one assumes your stock shows up all at once, and the other assumes it trickles in while you're still using it.
Economic Order Quantity (EOQ) and Economic Production Quantity (EPQ) are the two classic lot-sizing models every industrial engineer learns, and they get confused constantly because the formulas look almost identical and both answer the same question: "how much should I order or produce at a time to minimize total annual inventory cost?" The difference that actually matters is a single assumption about how replenishment arrives— instantaneously, or gradually while demand keeps eating into it. Get that assumption wrong and you'll size a production run using a formula built for a supplier delivery, and understate your required batch size.
EOQ (Economic Order Quantity) assumes the entire order quantity arrives in inventory in one instant — a truck backs up to the dock, the whole lot is unloaded, and inventory jumps straight up to its maximum level before demand starts drawing it back down. This is the right model for purchased items from an external supplier: EOQ = √(2DS / H), where D is annual demand, S is the fixed cost per order (setup/ordering cost), and H is the annual holding cost per unit.
EPQ (Economic Production Quantity), also called the Economic Manufacturing Quantity (EMQ), assumes the lot is produced internally at a finite production rate p, while demand at rate d is simultaneously drawing down the same stock during the production run. Inventory therefore ramps up gradually during production (at rate p − d, not all at once) and only starts falling once the production run ends. Because stock never has to cover the full lot size at any single moment, the maximum inventory level — and the resulting holding cost — is always lower than EOQ predicts for the same lot size, which is why EPQ intentionally recommends a larger optimal lot: EPQ = √[ (2DS / H) × (p / (p − d)) ].
If the answer is yes — a purchase order fills the warehouse in a single delivery — use EOQ. If the answer is no — a machine is actively producing units, and downstream demand is drawing from the same growing pile in real time — use EPQ. The tell is always the production rate p: EOQ has no such rate because replenishment is treated as instantaneous, while EPQ requires knowing both p and d to compute how fast inventory actually accumulates net of consumption. A common mistake is applying EOQ to an internally manufactured item anyway, which understates the optimal batch size and — because EOQ ignores the ramp-up period — also misstates the true holding cost and cycle time, since it assumes a maximum inventory level the process will never actually reach.
The formulas do look like close cousins, and mathematically EOQ is a special case of EPQ (as p → ∞). But treating EPQ as an afterthought correction misses the point: the entire inventory profile is different. EOQ's classic sawtooth chart — instant spike, straight-line decline to zero — simply does not describe what happens on a production line where the same machine is simultaneously filling and draining the same buffer. Using EOQ's peak-inventory assumption (Q) for a manufactured item overstates the actual maximum stock level, which cascades into an overstated holding-cost estimate and an understated optimal batch size — errors that compound directly into the total annual cost estimate the whole exercise exists to minimize.
Explains the real distinguishing assumption between Economic Order Quantity (EOQ) and Economic Production Quantity (EPQ) — instantaneous vs. gradual replenishment — using a side-by-side inventory-sawtooth comparison and the two lot-sizing formulas.
Both formulas answer the same optimization question (minimize total annual ordering/setup cost plus holding cost) and both produce a single-number "optimal lot size." Students often memorize EOQ = √(2DS/H) and EPQ = √(2DS/H · p/(p−d)) as if EPQ were just EOQ with a bolt-on correction term, without internalizing that the correction term exists because the two models describe physically different replenishment processes — one instantaneous, one gradual and concurrent with demand.
EOQ: Q* = √(2DS/H), assuming the full order quantity Q arrives in inventory at a single instant, after which only demand D reduces stock until it reaches zero and the next order is placed.
EPQ: Q* = √[(2DS/H) × (p/(p−d))], where inventory accumulates during the production run at the net rate (p − d) — production rate p minus demand rate d — because demand is being satisfied out of the same stock simultaneously. Production runs until Q units have been made, then stops while demand alone depletes the accumulated stock (which peaked at Q(1 − d/p), lower than Q itself) down to zero.
As p becomes very large relative to d, the factor p/(p−d) approaches 1 and the EPQ formula converges to the EOQ formula — confirming that EOQ is the limiting case of EPQ when production (or delivery) is effectively instantaneous.
The choice matters most in batch-manufacturing environments running multiple products on shared equipment, where a planner must decide production run sizes for internally made components (use EPQ) versus purchased raw materials or components sourced from suppliers (use EOQ). Getting it backward — applying EOQ logic to a production line — systematically underestimates the ideal batch size and understates the true carrying cost curve, which can drive excessive changeovers (too-frequent setups) chasing a lot size that was never actually optimal for a gradually-replenished process. This same instantaneous-vs-gradual distinction also underlies more advanced lot-sizing extensions, like EPQ models with planned shortages or EOQ with quantity discounts, both offered as calculator variants in this studio.
EOQ. Supplier deliveries are treated as arriving essentially all at once relative to the consumption timeline, matching EOQ's instantaneous-replenishment assumption.
EPQ. Because the same production line is creating inventory while demand simultaneously draws it down, the gradual-replenishment assumption in EPQ correctly captures the net accumulation rate (p − d) instead of treating the full batch as available instantly.
Yes. Because the correction factor p/(p−d) is always greater than 1 (since production rate p must exceed demand rate d for the process to keep up), EPQ's optimal lot size is always at least as large as EOQ's, and strictly larger whenever p is finite.
The factor p/(p−d) grows very large, driving the optimal lot size way up. This reflects a real physical constraint: if a process can barely keep pace with demand, it needs a much longer, larger production run to ever build meaningful buffer stock, since inventory accumulates only at the slim net rate (p − d).
Yes — both have well-known variants (EOQ with backorders, EPQ with backorders) that add a backorder cost term and shift some of the cycle into a planned shortage period, which can further reduce total annual cost when backordering is cheaper than holding extra safety stock, at the expense of occasional stockouts.
Try our Industrial & Systems Engineering Studio
More calculators, simulators, and guides for this discipline.
Industrial Engineering Reference (Full Access)
✨ Premium ContentA zoomable interactive reader — free preview, then unlock the full set.