GATE Production and Industrial Engineering (PI): Operations Research and Operations Management (linear programming, transportation, queuing, forecasting, EOQ and inventory, scheduling, CPM and PERT, break-even and NPV), previous-year questions solved, part 2
Part 2 of 2 of the Operations Research and Operations Management hub: 19 more official GATE PI questions on queuing models, forecasting and aggregate planning, inventory control: EOQ and quantity discounts, scheduling, sequencing and supply chain, project management: CPM and PERT and engineering economy: break-even, depreciation and NPV, each solved step by step and checked against the official answer key. Part 1 covers linear programming, transportation and network models.
Questions from official GATE PI papers (IISc Bengaluru 2024, IIT Roorkee 2025, IIT Guwahati 2026); solutions by Skill-Lync. · Updated 19 Nov 2026 · 19 of 26 questions on this page
Poisson arrivals at a single server queue: what distribution does the inter-arrival time follow?
- (A)Poisson distribution
- (B)Uniform distribution
- (C)Exponential distribution
- (D)Binomial distribution
Answer: (C) Official key: C
- If arrivals form a Poisson process of rate , the number of arrivals in a time is Poisson with mean . The probability of no arrival in time is .
- The next arrival is later than exactly when there is no arrival in , so . That is the exponential distribution, with density and mean .
- The first letter M in M/M/1 stands for Markovian, meaning exponential (memoryless) inter-arrival times, and the second M means exponential service times. The Poisson and binomial laws count events, so they describe the number of arrivals, not the gap between them. The uniform law has no memoryless property.
Check: we built a Poisson stream at rate 2 per unit time from 4 million tiny independent slots and measured 7946 gaps. The mean gap was 0.503 (theory 0.5), the standard deviation over the mean was 0.990 (exponential: 1, uniform: 0.577), the largest gap between the empirical and exponential distributions was 0.0088, and matched .
Call centres, tool cribs and repair counters are sized with M/M/1 and M/M/c models that assume random independent arrivals, which means exponential gaps between customers. Before using the model, analysts time a few hundred real arrivals and check that the gaps have a standard deviation close to their mean.
M/M/1 queue, arrivals 3 per minute, average number in system n: average time in system
Which ONE of the following is the average time a customer spends in the system (in minutes) at steady state?
- (A)
- (B)
- (C)
- (D)
Answer: (B) Official key: B
- Principle: Little's law links the average number in a system , the arrival rate and the average time in the system by . It holds for any stable queue, with no assumption about the service-time distribution.
- Here customers and customers per minute (a utilisation below 1 means the queue is stable and a steady state exists).
- minutes. Option (B).
Trap: multiplying instead of dividing (option A), or bringing in a service rate that the question does not give (option C looks like dividing by ). Checking against the M/M/1 formulas, and , gives . We also ran a seeded event simulation of 200000 customers at per minute. For () the time-average number in the system was 1.006 (formula 1.000) and the mean time in the system 0.3362 min, against min. For () the numbers were (formula 4.00) and min against min. Only (B) matched at both utilisations.
Little's law is used on factory floors to convert a count of jobs waiting in a cell, which is easy to see, into a lead time, which is what the customer feels: lead time equals work in process divided by throughput. The same relation sizes buffers in assembly lines and queues at service counters.
Which forecasting method needs the least data: econometric, regression, ARIMA or simple exponential smoothing?
- (A)Econometric forecasting method
- (B)Linear regression method
- (C)ARIMA method
- (D)Simple exponential smoothing method
Answer: (D) Official key: D
- Simple exponential smoothing updates one number: . The old forecast already carries the whole history, so the method can start from a single observation () and afterwards needs only the latest demand and the latest forecast.
- Linear regression has to estimate a slope and an intercept from a set of points, and needs at least three to leave one degree of freedom for judging the fit (more to be reliable).
- An econometric (causal) model needs a history of the demand and of every explanatory variable, such as price, income or advertising, and has more parameters to estimate.
- ARIMA (Box-Jenkins) models need a long, stationary series, usually 50 or more observations, to identify and estimate the model. So (D) uses the least data.
Check: a forecast built from the full 40-period history agreed with the forecast rebuilt from only the previous forecast and the latest demand (difference below ). Minimum useful observations: simple exponential smoothing 1, regression 3, econometric with two causal variables 4, ARIMA about 50 (the last is a guideline, not a theorem).
Spare-parts and consumables planners run simple exponential smoothing on thousands of items because a new part with a few weeks of history can still be forecast. Regression and ARIMA are kept for high-volume items with long histories where the extra data pays for itself.
Linear regression of annual profit on marketing expenditure: estimated profit at an expenditure of 5 units
| Year | Expenditure for marketing (units) | Annual profit (units) |
|---|---|---|
| 1 | 3 | 22 |
| 2 | 4 | 27 |
| 3 | 6 | 36 |
Answer: 31.43 units Official key: 30.00 to 33.00
- Causal forecasting by simple linear regression: fit by least squares, with the marketing expenditure and the profit, then read the forecast at the new . The fit is a straight line through the point of means.
- Means: and .
- Sums of squares about the means: and .
- Slope profit units per expenditure unit. Intercept units.
- At : 31.43 units.
Trap: forecasting from the average ratio profit/expenditure, which gives units (outside the key range), or reading the answer off the 3-year profit trend alone, ignoring the marketing spend. We solved the normal equations with exact fractions, cross-checked with and with a brute-force search over minimising the squared error: all three agree.
Regression of sales or profit on a driver such as advertising spend, price or housing starts is the standard causal forecast behind budget and capacity plans. Planners use it only inside the range of the data, since a line fitted to three years says little about spending levels never tried.
Aggregate planning for toy cars: inventory at the end of November from the partial optimal plan
| Month | September | October | November | December |
|---|---|---|---|---|
| Demand (in units) | 8000 | 6500 | 10500 | 9500 |
The organization has ONLY the following options available for aggregate planning to meet the demand: (1) in-house production, (2) subcontracting, and (3) inventory holding. Assume that the quantity subcontracted in a given month is available in the same month.
A part of their optimal solution is as follows:
Inventory at the end of October = 1500 units
Units produced in-house in November = 8000 units
Units subcontracted in November = 1000 units
The inventory at the end of November is ______ units (in integer).
Answer: 0 units Official key: 0 to 0
- Principle: in aggregate planning, stock carries from month to month by a balance: closing inventory = opening inventory + in-house production + subcontracted units demand. Subcontracted units arrive in the same month, so they count in that month.
- November: opening inventory is the October closing stock of 1500 units. Supply in the month is units.
- November demand is 10500 units, so closing inventory 0 units.
Check: the supply exactly equals demand, so none is left over, and it is consistent with an optimal plan, since carrying stock into December would only add holding cost when November's own production and subcontracting already cover demand. A key of 0 is a real answer here, not a missing key. Trap: forgetting the opening stock of 1500 units and computing , which is impossible for stock, or adding the demand of the next month. The balance in code gives 0 units (supply 10500 against demand 10500).
Planners of seasonal products such as toys build stock ahead of peaks and subcontract the overflow, and the month-by-month stock balance is the first check on any such plan. The inventory carried is what links one month's decisions to the next.
EOQ: ordering cost per order up 20% and demand up 20%: percentage change in time between orders?
Answer: 0 % Official key: 0 to 0
- The classical economic order quantity is , with demand per unit time, ordering cost per order and holding cost per unit per unit time.
- The time between orders is . It depends on the ratio .
- Now and , so is unchanged and is unchanged: the change is 0 %.
- The order quantity itself does rise, by the same factor , but so does the number of units used per unit time, so the gap between orders stays the same.
Check: with , INR, INR per unit per year, a total-cost scan over gave units and years. With and INR it gave units and years, the same. Traps: reporting the 20 % rise in order quantity, or changing only (cycle time would fall by 8.7 %).
Purchasing teams recompute the EOQ and the order interval whenever demand or the cost of placing an order changes, because the replenishment calendar follows the interval. When both drift up together, as when volume and administration cost grow in step, the ordering rhythm stays the same and only the lot size grows.
Economic order quantity of bulbs: demand 2000 per year, INR 300 per order, INR 30 carrying per bulb per year
Answer: 200 bulbs Official key: 199 to 201
- EOQ model: total yearly cost is ordering cost plus carrying cost, . It is lowest where the two parts are equal, which gives .
- Here bulbs per year, INR per order and INR per bulb per year (the carrying cost is already per bulb, so no percentage is needed).
- 200 bulbs.
- Check: at 200 bulbs the orders cost INR and carrying costs INR, equal, total 6,000 INR. Ordering 150 gives 6,250 INR and 250 gives 6,150 INR, both dearer.
Trap: forgetting the factor 2 in the numerator, or putting the number of orders (10 per year) in place of the order size. We scanned from 1 to 1000 for the lowest total cost and it is 200.
EOQ is the first calculation behind reorder sizes for consumables such as lamps, fasteners and cutting inserts in a plant stores. It balances the cost of raising a purchase order against the cost of capital and space tied up in stock, and it is deliberately flat near the optimum so a rounded pack size costs very little.
Optimal order size with price breaks at 1200 and 2000: demand 5000, INR 400 per order
| Quantity of item ( in units) | Unit price of item (Rupees) |
|---|---|
| 10 | |
| 8 | |
| 7 |
Answer: 2000 units Official key: 1995 to 2005
- All-units quantity discount: compute the EOQ at each price, with carrying cost and . If it falls outside its own price band, move it to the nearest band edge. Then compare total cost (purchase + ordering + carrying) across the candidates and take the cheapest.
- At INR 10: , inside to , valid. At INR 8: , inside to , valid. At INR 7: , below the 2000 needed for that price, not valid, so the candidate becomes .
- Total cost . At , price 10: about INR 53,464. At , price 8: about INR 43,098. At , price 7: INR 38,100.
- The lowest is INR 38,100, so the optimal order size is 2000 units.
Trap: stopping at the EOQ of the cheapest price (1,380 units, which is not allowed at that price) or at the first valid EOQ (1,291 units at INR 8). The saving of INR 1 per unit on 5000 units outweighs the extra holding cost, so the bigger order wins. We scanned every from 1 to 10,000 under the price table: the total cost is lowest at 2000.
Suppliers of fasteners, steel and packaging quote price breaks, and buyers have to weigh the unit saving against the cost of holding more stock. Checking whether each EOQ is allowed at its own price, then comparing total costs, is how purchasing decides whether to take the bulk tier.
Which ONE of the options is NOT an assumption of the economic order quantity (EOQ) model
- (A)Demand is known and deterministic.
- (B)Lead time is zero.
- (C)Stockouts are permissible.
- (D)Holding cost for the quantity in stock is linearly increasing with the order quantity.
Answer: (C) Official key: C
- Principle: the basic EOQ model balances ordering cost against holding cost with . That formula is valid only when demand is steady and known, the whole order arrives at once when stock reaches the reorder point, and no shortage is ever allowed.
- (A) Demand known and deterministic: an assumption. Without it the stock does not fall in a straight line and there is no single sawtooth cycle to cost.
- (B) Lead time is zero: an assumption in the textbook statement (instantaneous replenishment). A known constant lead time can be allowed, but it only moves the reorder point earlier by demand during lead time and leaves unchanged, so the model treats lead time as zero or fixed.
- (D) Holding cost linear in the quantity held: an assumption. The yearly holding cost is , proportional to the average stock, which is what makes the cost curve and gives the square-root formula.
- (C) Stockouts are permissible: the opposite. The basic model assumes shortages are not allowed. If shortages are permitted at a cost, a different model applies. So the statement that is NOT an assumption is (C).
Check with numbers: for units/year, INR/order, INR/unit/year, a scan over puts the cheapest order quantity at 1095 units, matching , for a yearly cost of 2191 INR. If planned shortages are allowed at 8 INR per unit short per year, the best order quantity moves to 1225 units and the yearly cost drops to 1960 INR: the EOQ formula is then no longer the optimum, so allowing stockouts changes the model. A daily simulation with a 1 day or 9 day lead time found the same best order quantity (1100 and 1100 units on a 100-unit grid), so lead time does not break the formula.
Repeat purchases of standard stock, such as fasteners, packaging or consumables, are ordered with EOQ-type logic. When a stockout would stop a production line, buyers keep a safety stock on top of it and do not rely on allowing shortages, which is why the EOQ assumption of no stockouts is usually kept in practice.
Which of location, inventory, distribution and machine scheduling are supply chain decisions?
- (A)Location
- (B)Inventory
- (C)Distribution
- (D)Machine scheduling
Answer: (A) and (B) and (C) Official key: A;B;C
- A supply chain covers the flow of material and information from suppliers through plants and warehouses to customers. Its standard decision areas are location (where to put plants, warehouses and suppliers), inventory (how much to hold, and where, to buffer the flow), transportation and distribution (modes, routes and network), plus sourcing, pricing and information sharing.
- (A) Location, (B) Inventory and (C) Distribution are all supply chain decisions.
- (D) Machine scheduling sequences jobs on machines inside one plant. It is a production planning and control decision, a shop-floor input to the chain rather than a design decision about the chain itself, so it is not selected.
Check: the keyed answer is the set A, B and C. Do not tick (D) just because scheduling affects delivery times.
A car maker choosing where to build a new plant, how many days of stock to hold at a regional warehouse, and whether to ship by rail or road is making supply chain decisions. Deciding which job runs next on a machining centre is handled by the plant scheduling team.
Five jobs A to E (9, 6, 4, 5, 8 days) on one machine by the SPT rule: average flow time in days?
| Job | A | B | C | D | E |
|---|---|---|---|---|---|
| Processing time (in days) | 9 | 6 | 4 | 5 | 8 |
Answer: 16.6 days Official key: 16.6 to 16.6
- The SPT rule runs the job with the shortest processing time first. Order: C (4), D (5), B (6), E (8), A (9).
- All jobs are available at , so the flow time of a job equals its completion time: days.
- Average flow time 16.6 days.
Check: we tried all 120 sequences and none gave a lower mean flow time than SPT, which is the known optimum for mean flow time on a single machine. Trap: processing in the order A, B, C, D, E gives 19.8 days, and forgetting to divide by 5 gives 83.
Dispatchers on a single bottleneck machine, such as a heat-treatment furnace or a CNC router, use shortest-processing-time dispatching to cut average work-in-process and waiting time. Its weakness is that long jobs can be pushed back repeatedly, so shops often cap waiting time.
Four programs on one server with the shortest processing time (SPT) rule: number of tardy jobs
| Program | 1 | 2 | 3 | 4 |
|---|---|---|---|---|
| Processing time (in minutes) | 60 | 40 | 90 | 100 |
| Promised time (i.e., due date) | 12 noon | 11:30 am | 1:00 pm | 1:30 pm |
If the CC follows shortest processing time (SPT) rule to sequence the computer programs, then number of tardy jobs is ______.
Note: Ignore program change-over times.
Answer: 2 jobs Official key: 2 to 2
- Rule: SPT sorts the jobs by processing time, shortest first, and runs them one after another with no gaps. A job is tardy if its completion time is later than its due date.
- Sequence by processing time: program 2 (40 minutes), program 1 (60), program 3 (90), program 4 (100). Measure time from 10:00 am.
- Program 2 finishes at 40 minutes (10:40 am), due 11:30 am: on time. Program 1 finishes at 100 minutes (11:40 am), due 12 noon: on time.
- Program 3 finishes at 190 minutes (1:10 pm), due 1:00 pm: 10 minutes late, tardy. Program 4 finishes at 290 minutes (2:50 pm), due 1:30 pm: 80 minutes late, tardy.
- The number of tardy jobs is 2.
| Program (SPT order) | Processing (min) | Completion (min after 10:00) | Due (min after 10:00) | Lateness (min) |
|---|---|---|---|---|
| 2 | 40 | 40 | 90 | 0 |
| 1 | 60 | 100 | 120 | 0 |
| 3 | 90 | 190 | 180 | 10 |
| 4 | 100 | 290 | 210 | 80 |
Trap: converting the due times wrongly, for example reading 1:00 pm as 100 minutes, or measuring from midnight. Due dates in minutes after 10:00 are 120, 90, 180 and 210. We ran the SPT schedule in code and got 2 tardy jobs. As a side check, all 24 sequences were tried: the fewest tardy jobs possible is 1, so SPT happens to be the best here, though SPT is only guaranteed to minimise mean flow time (the earliest due date rule gives 2 tardy jobs).
Dispatching by shortest processing time is the default rule on many job shops and print or machining queues because it cuts the average waiting time and keeps the work in process low. Its weakness is that long jobs wait longest and may miss their due dates, so shops often combine it with an earliest-due-date check.
Six activities, A, B, C start; D, F need A and B; E needs B and C: minimum dummies in an AOA network?
| Activity | Precedent activities |
|---|---|
| A | None |
| B | None |
| C | None |
| D | A, B |
| E | B, C |
| F | A, B |
- (A)0
- (B)1
- (C)2
- (D)3
Answer: (D) Official key: D
- In an activity-on-arrow diagram each activity is an arrow between two events, an event means all arrows into it are finished, and no two arrows may join the same pair of events. A dummy (zero-duration) arrow is added to show a precedence the real arrows cannot, or to separate two parallel arrows.
- D and F have the same predecessors A and B, so they start from the same event X, and A and B must both finish into X. E needs B and C but not A, so it starts from a different event Y, and B must also finish into Y. B is the shared activity and cannot end at both X and Y without help, so B ends at its own event and two dummies carry it to X and to Y: dummies 1 and 2.
- D and F both run from X and both end the project. Two arrows between the same two events are not allowed, so F ends at an extra event and a third dummy joins it to the finish event: dummy 3.
- Three dummies are needed. With no dummy, A and B would both run from the start to X, which is not allowed. The check below shows 1 and 2 dummies do not work either. Answer (D).
Check: we searched every activity-on-arrow network with up to 6 events and 0, 1 or 2 dummies, with one start and one end event, and none reproduced the table. The 3-dummy network (A 0-2, B 0-1, C 0-3, D 2-5, E 3-5, F 2-4, dummies 1-2, 1-3 and 4-5) passed a reachability test of every predecessor set, and removing any one dummy broke it. Trap: counting only the first two dummies and forgetting the one that separates D from F gives 2.
Project schedulers draw dummy arrows (logic links) in construction and shutdown planning so that each task depends on exactly the right earlier tasks. Modern scheduling software uses activity-on-node networks, which need no dummies, but AOA diagrams are still met in older plans and in exams.
Critical path of a project network with eight activities A to H (durations in days)
- (A)1–2–3–6
- (B)1–4–3–6
- (C)1–5–6
- (D)1–4–5–6
Answer: (B) Official key: B
- Critical path method: the critical path is the longest path from the start node to the end node. It sets the project duration, and every activity on it has zero float (no slack).
- List the paths from node 1 to node 6 and add durations. (A, D, G): days. (B, E, G): days. (C, H): days. (B, F, H): days.
- The longest is 1–4–3–6 at 15 days, activities B, E and G. Option (B).
- Check with a forward and backward pass: the earliest time of node 3 is , and of node 6 is days. Floats in days: A 1, D 1, C 9, F 10, H 9, while B, E and G have float 0.
Trap: picking the path with the most activities or the one that looks most direct. 1–2–3–6 comes within a day (14 days) and is the near miss. We enumerated every path with its length (1-4-3-6 = 15 days; 1-2-3-6 = 14 days; 1-5-6 = 6 days; 1-4-5-6 = 5 days) and ran the passes in code to confirm the zero-float activities are exactly B, E and G.
The critical path tells a project planner which tasks decide the finish date, for example in a plant shutdown, a die changeover or a line installation. Effort and expediting money go on those tasks, while tasks with float can be scheduled around the available crews and cranes.
Painting activity of 6 days with total float and free float both zero: which statement is true
Which ONE of the following is TRUE for the painting activity?
- (A)Painting is a dummy activity.
- (B)Painting is a non-critical activity.
- (C)The start of painting activity can be delayed without causing delay in completion of the construction project.
- (D)The start of painting activity cannot be delayed without causing delay in completion of the construction project.
Answer: (D) Official key: D
- Definitions: total float is the time an activity can be delayed without delaying the project end, . Free float is the delay possible without delaying the earliest start of any following activity. Free float can never exceed total float.
- Total float zero means : there is no slack between the earliest and latest start, so the activity is on a critical path. That rules out (B).
- A dummy activity has zero duration and is used only to show a dependency. Painting takes 6 days, so (A) is false.
- With zero total float, any delay to the start moves the finish of the activity and every critical successor with it, so the project end date slips by the same amount. (C) is false and (D) is true.
Trap: the free float figure adds nothing once the total float is zero, because . The reverse is not true: an activity can have free float zero but a positive total float (it can hold up a successor but not the project end). We checked the rule on 400 random networks: all 1524 activities with zero total float also had zero free float, and delaying the start of each by one day lengthened the project every time (276 other activities had zero free float with positive total float). In a worked example (excavate 4, foundation 7, walls 8, services 5, painting 6, handover 2 days) painting has , , , and starting it one day late moves the project from 27 to 28 days.
On a site programme, the zero-float activities are the ones the planner watches daily, because a late painting crew means a late handover. Activities with positive float are used to level labour and equipment without moving the finish date.
Machines A (INR 20,000, 10 years) and B (INR 15,000, 20 years), salvage INR 5,000: book value gap at year 5?
| Machine A | Machine B | |
|---|---|---|
| Purchase price | INR 20,000 | INR 15,000 |
| Estimated useful life | 10 years | 20 years |
| Estimated salvage value | INR 5,000 | INR 5,000 |
Answer: INR 0 Official key: 0 to 0
- Straight-line depreciation writes off the same amount every year: . The value (book value) after years is the purchase price minus .
- Machine A: . After 5 years: .
- Machine B: . After 5 years: .
- The difference is 0 INR. The key is exactly 0 (a real answer, not a missing one).
Check: we wrote out the year-by-year book values for both machines and both ended year 5 at INR 12,500. Trap: ignoring the salvage value and writing off the whole price gives book values INR 10,000 and 11,250, a gap of INR 1250.
Fixed-asset registers in factories apply straight-line depreciation to machines, tooling and vehicles over their useful life, and the book value after a given year feeds replacement and sale decisions. Two assets with very different lives and prices can still carry the same book value on a given date, as here.
Make versus buy for toy batteries: annual demand at which to switch from buying to making
Answer: 10000 batteries per year Official key: 9995 to 10005
- Make-versus-buy break-even: the two options cost the same at the demand where the extra fixed cost of making is exactly paid back by the saving per unit.
- Cost of making: INR per year. Cost of buying: INR per year, where is the annual demand in batteries.
- Set them equal: , so and 10,000 batteries per year.
- Direction check: at 5,000 batteries buying costs INR 450,000 against INR 550,000 for making, so buy; at 15,000 buying costs INR 1,350,000 against INR 1,250,000, so make. At 10,000 both cost INR 900,000.
Trap: dividing by the variable cost (2,00,000 / 70 = 2,857) or by the buying price (2,00,000 / 90 = 2,222) instead of by the saving of INR 20 per unit. We scanned annual demand from 1 to 20,000 and the first demand at which making is no dearer than buying is 10,000.
Make-or-buy decisions for components such as small motors, castings and packaging turn on exactly this crossover: a fixed tooling or plant cost against a lower cost per piece. Below the break-even volume the supplier's price wins, above it in-house production wins, and the answer shifts whenever either price changes.
Net present value of INR 2,00,000 invested for INR 1,00,000 a year for three years at 15% per year
The net present value (NPV) of his/her investment is INR ______ (rounded off to the nearest integer).
Answer: INR 28323 Official key: 28319 to 28324
- Principle: net present value is the sum of all cash flows, each divided by , where is the discount rate and the year. Money received later is worth less today.
- The investment of INR 2,00,000 is at year 0, so it is a negative flow of and is not discounted.
- The three payouts are discounted separately: , and . Their sum is INR, where is the present worth factor of a three-year annuity at 15%.
- INR, which is INR 28323 to the nearest integer.
Trap: adding the payouts without discounting gives INR, which looks profitable by far more than it is. The positive NPV means the investment earns more than the 15% required return. We summed the discounted flows in code over years 0 to 3 and got 28322.51 INR.
Companies compare a machine purchase, a software licence or a plant upgrade by the NPV of its savings against its cost, using the company's required return as the discount rate. A positive NPV means the project clears that return.
Break-even quantity when 5,000 units earn a profit of 20% of sales revenue (fixed cost INR 3,00,000)
The break-even quantity for that product is _____ units (rounded off to the nearest integer).
Answer: 2667 units Official key: 2665 to 2679
- Principle: break-even is where revenue equals total cost, so , with fixed cost , selling price per unit and variable cost per unit. The selling price is not given, so it must be found from the profit information first.
- Total cost at 5000 units: INR. Profit is 20% of revenue , so cost is 80% of revenue: , giving INR.
- Selling price INR per unit.
- , which is 2667 units to the nearest integer.
Trap: taking 20% of revenue as 20% of cost, or as a mark-up on variable cost. The profit is stated on sales revenue, so cost is 80% of revenue. Check in code: profit at 5000 units is 20.0% of revenue; at 2666 units the result is -75 INR (a small loss) and at 2667 units it is 38 INR (a small profit), so the break-even point lies between them at 2666.67 units.
Break-even quantity tells a plant how many units must sell to cover its fixed overheads before any profit appears, and it is recalculated whenever the price, a material cost or a fixed cost such as rent or a new machine changes. Managers read the margin of safety, the gap between actual sales and this quantity, as a measure of risk.