Skip to content
Tutorials Center
My learning

Topics

Lesson 46 of 49 ยท Projects and Practice

Practice Problems for Chapters 5 to 8

Fourteen worked practice problems on EOQ, EPQ, newsvendor, scheduling, queues, decisions, games and heuristics, with a self-check program and an answer table.

You learn operations research by doing it. This set has fourteen problems covering the four chapters of this part of the course. Try each one on paper first, then check yourself with the program in the middle of the lesson and the answer table at the end. If an answer differs from yours, find out why before moving on: the mistake is usually a unit (per year versus per day) or a formula that was used outside its assumptions.

A map of six practice topics
Caption: The practice set follows the course: inventory, scheduling, queues, simulation, decisions and games, and heuristics.

What you will learn

  • To apply each model of the course to a fresh numerical problem.
  • To check answers with short programs.
  • To find your own weak spots before an exam or a project.

How to use this set

  1. Read the problem and write down the model you will use (EOQ? M/M/c? payoff table?).
  2. List the given quantities with their units.
  3. Solve by hand with a calculator, then run the self-check.
  4. If you are stuck, reread the relevant lesson: inventory in chapter 5, queues in chapter 6, decisions and games in chapter 7, heuristics in chapter 8.

Section A: Inventory

Problem 1 (EOQ). A clinic uses 5,000 boxes of gloves a year. Each order costs $80 to place and holding a box costs $4 per year. Find the order quantity, the number of orders per year and the yearly relevant cost.

Problem 2 (EPQ). A plant makes 19,200 units a year at full speed and sells 4,800 a year. Each production run costs $150 to set up and holding costs $3 per unit per year. Find the run size, the peak inventory and the yearly relevant cost.

Problem 3 (newsvendor). A caterer buys a seasonal dessert for $12, sells it for $30 and salvages leftovers for $4. Demand is normal with mean 120 and standard deviation 25. How many should she buy?

Problem 4 (reorder point). Weekly demand has mean 80 and standard deviation 15 (independent weeks). The lead time is 3 weeks. Find the reorder point for a 97.5 percent cycle service level (z = 1.96).

Section B: Scheduling

Problem 5 (SPT). Four jobs with processing times 4, 2, 6 and 3 hours are waiting for one machine. Find the order that minimizes the mean completion time and that mean.

Problem 6 (Johnson's rule). Four jobs go through machine 1 and then machine 2. Times (M1, M2): J1 (5, 2), J2 (1, 6), J3 (7, 7), J4 (3, 4). Find the sequence and the makespan.

Section C: Queues

Problem 7 (M/M/1). A repair desk gets 6 customers an hour; the technician serves 8 an hour. Find ฯ, L, Lq, W and Wq (in minutes where useful).

Problem 8 (Little's law). On average, 12 orders are being processed in a workshop and 3 orders arrive per hour. How long does an order spend in the workshop?

Problem 9 (Erlang C). A phone line gets 12 calls an hour and each of 3 agents handles 5 calls an hour. Find the probability that a caller waits and the average wait in line.

Section D: Decisions and games

Problem 10 (EMV and EVPI). Option A pays 100 or 40 and option B pays 70 or 60 in two equally likely states. Find the best EMV and the EVPI.

Problem 11 (zero-sum game). The row player's payoff matrix is [[2, 5], [4, 1]]. Find the optimal mixed strategies and the value.

Problem 12 (Bayes). A disease affects 2 percent of a population. A screening test is positive for 95 percent of those with the disease and for 5 percent of those without it. What is the probability of having the disease after a positive result? (A teaching example of updating, not medical advice.)

Section E: Heuristics

Problem 13 (annealing). In simulated annealing a move worsens the cost by 3 at temperature 2. What is the acceptance probability?

Problem 14 (gap). A heuristic finds a route of 5,210 km; the best known route is 5,000 km. What is the optimality gap?

Self-check

The program solves all fourteen problems. Run it only after you have your own answers.

Python Check your answers

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

A practice generator

Try unlimited EOQ drills. Each press creates a new problem and checks your answer within 1 percent.

EOQ drill

Answers

Problem Answer
1 Q* โ‰ˆ 447 boxes, about 11.2 orders per year, cost about $1,789
2 Run size 800, peak inventory 600, cost $1,800
3 Critical ratio 18/26 = 0.692, so z โ‰ˆ 0.50 and Q* โ‰ˆ 132.6, about 133 desserts
4 Lead-time demand 240, standard deviation 26.0, safety stock about 51, reorder point about 291
5 Order 2, 3, 4, 6; completions 2, 5, 9, 15; mean 7.75 hours
6 Sequence J2, J4, J3, J1; makespan 20
7 ฯ = 0.75, L = 3, Lq = 2.25, W = 30 min, Wq = 22.5 min
8 W = L/ฮป = 4 hours
9 Probability of waiting about 0.647; average wait about 12.9 minutes
10 EMV(A) = 70, EMV(B) = 65, best 70; with perfect information 80, so EVPI = 10
11 Row plays row 1 with probability 0.5, column plays column 1 with probability 2/3, value 3
12 About 0.279 (positives are mostly false alarms because the disease is rare)
13 e^{โˆ’1.5} โ‰ˆ 0.223
14 Gap 4.2%

A worked solution: problem 9

Problem 9 asks for the Erlang C probability for ฮป = 12, ฮผ = 5, c = 3. Here is the full path by hand.

  1. Offered load: a = 12/5 = 2.4 Erlangs. Utilization ฯ = a/c = 0.8, so the system is stable.
  2. The sum over k = 0, 1, 2: 1 + 2.4 + 2.4ยฒ/2 = 1 + 2.4 + 2.88 = 6.28.
  3. The waiting term: (2.4ยณ/3!)/(1 โˆ’ 0.8) = (13.824/6)/0.2 = 2.304/0.2 = 11.52.
  4. C = 11.52 / (6.28 + 11.52) = 11.52/17.8 โ‰ˆ 0.647.
  5. Average wait in line: Wq = C/(cฮผ โˆ’ ฮป) = 0.647/(15 โˆ’ 12) โ‰ˆ 0.2157 hours, about 12.9 minutes.

Every step uses a formula from chapter 6, and each intermediate number can be sanity-checked: the probability is between 0 and 1, and the wait is positive and plausible for 80 percent utilization.

Study strategy

  • Translate first. Write the model name and the given symbols before computing anything. Most errors come from choosing the wrong model, not from arithmetic.
  • Estimate before you compute. For Problem 7, utilization 0.75 means a queue of a few customers, not hundreds. A wrong order of magnitude reveals an error at once.
  • Check units. Hours versus minutes and years versus days cause more wrong answers than any other slip.
  • Redo the missed ones. After checking, close the page and solve the problems you got wrong again from scratch a day later.
  • Make up variations. Change a number and predict how the answer should move (does doubling S double the EOQ?). If you can predict the direction and rough size, you understand the model.

When your answer differs

If your result does not match the table, do not just copy the right number. First compare the intermediate values: for an inventory problem, check D, S and H and their units; for a queue, check ฯ before anything else, because every other formula uses it. Next, ask whether you used the right model: Problem 2 looks like Problem 1 but needs the production factor 1 โˆ’ D/P, and Problem 11 needs the mixed-strategy formula only because there is no saddle point. Finally, test your answer for reasonableness by changing one input slightly and seeing whether the output moves in the direction and size you expect.

Common misconceptions

"Mixing units is harmless." Holding cost per year with demand per day gives nonsense. Convert first.

"A positive test means a high chance of disease." The prior matters: for a rare condition, even an accurate test gives many false positives (problem 12).

Check your understanding

Quick check ยท 4 questions

of 4 answered You scored / 4 ยท perfect!

  1. 1. D = 10,000, S = $50, H = $2.5. What is the EOQ?

    Q* = โˆš(2 ร— 10,000 ร— 50 / 2.5) = โˆš400,000 โ‰ˆ 632.
  2. 2. Which single quantity does Little's law connect to the average number in the system?

    L = ฮปW.
  3. 3. In problem 11 the row player's value is 3 because...

    With p = 0.5 the expected payoff is 3 against either column.
  4. 4. Which tool would you choose for a one-season order under random demand?

    A single selling season with random demand and different over- and under-stocking costs is the newsvendor problem.

Summary

  • Work each problem by model: identify it, list the units, solve and check.
  • Self-checks with short programs catch arithmetic slips quickly.
  • The hardest mistakes are unit mismatches and using a formula outside its assumptions.
  • Chapter 9 ends with a glossary, a cheat sheet and ideas for further study.

Next: Glossary