MGST391 BUSINESS ANALYTICS Illustrative Exams & Solutions Mid-Term & Final 2020-2021 MGST391 Business Analytics Illustrative Mid-Term Exam Questions This set of questions is illustrative of the type of questions to expect on the mid-term exam. As such, this set of questions may not include all the aspects of the course which are covered on the mid-term. This practice exam is meant to convey exam style and structure rather than content. Business Analytics and Excel Modeling 1. Which of the following is FALSE about modelling? (a) (b) (c) (d) 2. Which of the following is not part of the Business Analytics discipline? (a) (b) (c) (d) 3. Raw data is of little use by itself; you need to model to make sense of it The final quality of the analysis improves poor data. Modelling can help us to understand the Business process Counter-intuitive results can be very valuable. Excel Solver Simulation Forecasting Linear Programming Which of the following Excel functions would be best to determine the distribution of assigned grades in a course? (a) (b) (c) (d) VLOOKUP Exact VLOOKUP Range COUNTIF IF MGST391 Illustrative Exams & Solutions Page 1 4. A feasible solution to a linear programming problem: (a) (b) (c) (d) (e) 5. must satisfy all of the problem's constraints simultaneously need not satisfy all of the constraints, only some of them must give the minimum possible cost must give the maximum possible profit None of the above is true Infeasibility in a linear programming problem occurs when: (a) (b) (c) (d) (e) there is an infinite solution set a constraint is not needed more than one solution is optimal there is no solution that satisfies all of the constraints None of the above is true 6. Which of the following is TRUE? (a) (b) (c) (d) Re-Solving an LP problem in Solver will yield the same change to the objective function as using sensitivity analysis. Defining integer constraints avoids non-linearity errors. Iso-profit lines must be parallel to the constraint furthest from the origin. Non-binding constraints should be removed because they are redundant. 7. The allowable increase and decrease for constraints: (a) (b) (c) (d) 8. tells us the change of the right-hand side of a constraint that may be added or subtracted without changing the profit (objective function value) tells us the change that may be made in the right-hand side of a constraint before the shadow price changes is shown on the sensitivity output of integer programming problems all of the above are correct A shadow price is: (a) (b) (c) (d) the coefficient of a changing cell the value of one additional unit of a resource the maximum change for a resource always positive for maximization problems MGST391 Illustrative Exams & Solutions Page 2 USE THE FOLLOWING INFORMATION TO ANSWER THE NEXT FIVE INDEPENDENT QUESTIONS 9. How many of the constraints are binding? (a) (b) (c) (d) 10. 0 1 2 3 Suppose we are required to increase our contract for A to 6 units. How will this impact our profit? (a) (b) (c) (d) No impact Increases the profit by $10.00 Decrease the profit by $10.00 Cannot tell MGST391 Illustrative Exams & Solutions Page 3 11. Suppose we were offered a 50 kg pallet of material for $350. You must take it all. Should we take it? (a) (b) (c) 12. Suppose we could increase our available labour hours to 195 hours. What would our new profit level be? (a) (b) (c) (d) 13. $750 (stay the same) $1073.75 at most $528, but can’t be determined further with the information given at least $972, but can’t be determined further with the information given Which of the following is TRUE about Solver for this problem? (a) (b) (c) (d) 14. No, don’t take it Yes, take it Cannot tell The reduced cost is 0 because one does not want any more of a non-binding constraint. This is an example of an integer problem. The complete optimal solution is to make 5 of Product A, 10 of Product B, and 15 of Product C. The model is linear. Let X1, X2 and X3 represent binary decision variable where a value of variable Xi equal to 1 represents the performance of project i. Which formulation includes the constraint that one can only perform project 1 or project 2 (but not both) when project 3 is performed? (a) (b) (c) (d) (e) X1 + X2 < X3 X1 + X2 < 2 X3 X1 < X3 and X2 < X3 Two of the above are correct None of the above are correct MGST391 Illustrative Exams & Solutions Page 4 THE NEXT QUESTION SHOULD BE ANSWERED USING THE FOLLOWING INFORMATION. Farmer Sally is raising ostriches to sell at market. She wants to feed the ostriches a diet that meets all nutritional requirements at minimum cost. The diet will consist of a mixture of the 3 types of feed available. The following table shows the nutrition in 1 kilogram of each type of feed, along with their costs, and the minimum daily requirement of each nutritional ingredient. Finally, she does not want more than 40% of the feed to be of type Feed X. We can develop an algebraic representation of this problem and solve it as a linear program. Let X represent the number of kilograms of Feed X, Y the number of kilograms of Feed Y, and Z the number of kilograms of Feed Z that Sally will decide to use for her mixture. Carbohydrates (g) Protein (g) Vitamin A (mg) Fat (g) Cost (in cents) 15. Kilogram of Feed X 50 15 10 14 Kilogram of Feed Y 20 70 25 12 Kilogram of Feed Z 30 40 5 7 21 42 36 Minimum daily requirement 60 45 35 15 (min) 35 (max) Formulate this problem as a Linear Programming model: MGST391 Illustrative Exams & Solutions Page 5 16. Mugsy’s Muffins is trying to decide how many batches of their Plain Bran (P), Raisin Bran (R), Banana Bran (B) and Date Bran (D) muffins to bake. They use bran, eggs, flour, sugar, baking soda, raisins, bananas, and dates to make their different bran muffins. For each batch of muffins, they use: 0.8 kg of bran, 2 eggs, 0.6 kg of flour, 0.25 kg of sugar, and 3 mL of baking soda. In addition, Raisin Bran requires ½ kilogram of raisins for each batch, Banana needs one banana for each batch, and Date requires ¼ kilogram of dates for each batch. In addition to the total ingredient limitations, the store will offer its selection of bran muffins with the following requirements: at least 4 batches of Banana Bran, no more than 2 batches of Date Bran, and no more than 30% of all the bran muffins can be with raisins or dates. Each batch of bran muffins sells for $9.00 regardless of type. Mugsy wants to maximize his profit. Ingredient Bran Eggs Flour Sugar Baking Soda Raisins Bananas Dates Cost $2.00/kg $0.10/egg $1.50/kg $1.00/kg $0.01/mL $1.80/kg $0.15/banana $2.40/kg Supply 200 kg 4 dozen 160 kg 40 kg 500 mL 20 kg 15 bananas 12 kg Note: although this question was used in a previous exam, the current question is much shorter. But still just as difficult, if not more so. MGST391 Illustrative Exams & Solutions Page 6 Formulate (but do not solve) this problem in standard Linear Programming format. Do not formulate this as an Excel Solver problem. MGST391 Illustrative Exams & Solutions Page 7 USE THE FOLLOWING INFORMATION TO ANSWER THE NEXT 4 QUESTIONS. 17. Which of the following statements about the Solver solution output is FALSE? (a) (b) (c) (d) 18. Produce 400 units of Product 1 at Plant A. All of the product is produced. There are 4 changing cells. Product 2 is made at both plants. Which of the plants has excess capacity? (a) (b) (c) Plant A Plant B neither 19. What would occur in the cost minimization problem if both constraints were made to be < ? (a) no change (b) infeasible (c) changing cell values would become 0 (d) unbounded 20. What value does the objective function coefficient for Product 1 at Plant B need to be (at least) in order to have this product scheduled at Plant B in the optimal solution? (a) (b) (c) 10.00 13.00 16.00 MGST391 Illustrative Exams & Solutions Page 8 21. Solve the following Linear Programming model using graphical methods. The Valley Wine Company produces two types of wines – Nectar and Red. The company has 48 tonnes of grapes this season. However, they only have 54 cubic meters of storage space for ageing and 32 hours of processing time. Solve the LP model to determine the optimal number of 100-liter batches of Nectar and Red to produce to maximize profit. MAXIMIZE Subject to 4N + 3R (Profit, $000) 4N 6N 4N + + + 6R 6R 2R N, R < < < > 48 54 32 0 (grapes, tonnes) (storage space, meters3) (processing time, hours) (non-negativity) RED Label the constraints and feasible region clearly. Show solution method (iso-profit for $24,000 or corner point). Explicitly state the optimal solution. NECTAR MGST391 Illustrative Exams & Solutions Page 9 USE THE FOLLOWING INFORMATION FOR THE NEXT FOUR QUESTIONS: The Poppa’s Pizza Parlour is a small restaurant catering to patrons with a European taste for pizza. One of its specialties is Poppa’s Prize pizza. The manager must forecast weekly demand for these special pizzas so that he can order the special pizza shells weekly. Recent demand has been as follows: Week of June 2 June 9 June 16 June 23 June 30 July 7 22. Number of Poppa’s Prize Pizzas 50 65 53 56 55 60 Using a four−week moving average, what is the forecast for the week of July 14? (a) (b) (c) (d) FJuly 14 < 55 55 < FJuly 14 < 60 60 < FJuly 14 < 65 FJuly 14 > 65 23. If a four-week weighted moving average were used, what would the forecast for the week of July 14 be? (The weights are 0.4, 0.3, 0.2, and 0.1, with the highest weight for the most recent period). (a) FJuly 14 < 55 (b) 55 < FJuly 14 < 60 (c) 60 < FJuly 14 < 65 (d) FJuly 14 > 65 24. If the forecast of sales for July 7 is 64 and α = 0.2, what is the simple exponential forecast for July 14? (a) FJuly 14 < 55 (b) 55 < FJuly 14 < 60 (c) 60 < FJuly 14 < 65 (d) FJuly 14 > 65 MGST391 Illustrative Exams & Solutions Page 10 25. Assume that a forecast of 56 was used for the 6 weeks. Calculate the following (use chart below): (a) Bias: (b) MAD: (c) Standard Error: Yt 50 Ft 56 65 56 53 56 56 56 55 56 60 56 Error = Yt -Ft |Error| (Error)2 Total 26. Which one of the following statements about forecasting is FALSE? (a) (b) (c) (d) In order to achieve the objective of developing a useful forecast from the information at hand, the forecaster must select the appropriate techniques. This choice sometimes involves a trade-off between forecast accuracy and cost. There are a number of general types of forecasting techniques used for demand forecasting, including time series analysis, causal methods, and qualitative techniques. Time series express the relationship between the factor to be forecasted and related factors, such as promotional campaigns, economic conditions, and competitor's actions. A time series is a list of repeated observations of a phenomenon, such as demand, arranged in the order in which they actually occurred. MGST391 Illustrative Exams & Solutions Page 11 27. A pet food retailer has experienced demands for wild birdseed over three years by trimester as indicated in the table below (in kilograms). Compute the forecasts for each trimester of Year 4. Determine the seasonal index (to 2 decimal places) and assume a growth rate of 4% of the last three years average for the trend component. Year 1 Jan-Apr 4010 May-Aug 2670 Sep-Dec 3070 2 4070 2590 3220 3 4050 2610 3410 Total Total Seasonal Index Forecast for Year 4 28. Which one of the following forecasting methods will generate accurate forecasts when demands have a consistent trend pattern? (a) (b) (c) (d) 29. Simple moving average method. Weighted moving average method. Single exponential smoothing method. None of the above. Which of the following statements about forecasting is FALSE? (a) (b) (c) (d) The method for incorporating a trend in an exponentially smoothed forecast is called single exponential smoothing. The sum of absolute forecast errors is useful in measuring the MAD in a forecast. The standard deviation and the mean absolute deviation measure the dispersion of forecast errors. A negative bias in a forecast indicates that the forecast is likely to be higher than the actual. MGST391 Illustrative Exams & Solutions Page 12 30. Which of the following is not usually a reason for incorrect forecasts? (a) (b) (c) (d) (e) Wrong model is chosen. Past data is not collected properly. Unexpected outside events. The past is not a reflection of what will happen in the future. All of the above are reasons for incorrect forecasts. Use this Information to answer the next two questions: Minnie has just come up with a new invention. Initially, she did not know what her demand would be, but has tracked its progress over the last number of weeks. She has asked you for some advice regarding the appropriateness of her forecast, shown in the spreadsheet output below. 31. What would be the best thing Minnie could do to improve her forecast? (a) (b) (c) (d) 32. verify that the data she has collected is correct change to a model that includes a seasonality component change the value of the smoothing constant in her current model change to a model that includes a trend component Which of the following would improve the way for Minnie to judge how good her forecasting model is for these data? (a) (b) (c) (d) do a sensitivity analysis to determine how much additional data is needed use a moving average model either set the first time period’s forecast equal to the actual demand, just test it with the last few time periods here, or test it against new data make the forecast more accurate by increasing the number of decimals by one MGST391 Illustrative Exams & Solutions Page 13 MGST391 Illustrative Mid-Term Exam SOLUTIONS 1. (b) ***The final quality of the analysis improves poor data. – it does not. Because poor data will result in incorrect or poor decisions even with analytics. 2. (a) ***Excel Solver - it is a modelling tool 3. (a) (b) (c) (d) VLOOKUP Exact (for matching IDs) VLOOKUP Range (for assigning grades) *** COUNTIF (if the grades are already assigned, want to count the occurrence of each letter grade) IF (could be done but would be cumbersome) 4. (a) (b) (c) (d) (e) *** Correct answer. The optimal solution must satisfy all the constraints. Not true in a profit maximization problem. Not true in a cost minimization problem. Wrong answer. 5. (a) (b) (c) (d) (e) This means that the solution is unbounded. This does not affect infeasibility. This means that there are multiple solutions not infeasible ones. *** Correct answer. Wrong answer. 6. (a) *** Re-Solving an LP problem in Solver will yield the same change to the objective function as using sensitivity analysis – TRUE Defining integer constraints avoids non-linearity errors – non-sensical statement. Iso-profit lines must be parallel to the constraint furthest from the origin – usually not parallel to constraints. Non-binding constraints should be removed from consideration because they are redundant. – False since non-binding constraints can become binding if a resource or objective function changes. (b) (c) (d) 7. (a) (b) (c) (d) is incorrect as RHS changes in scarce resources will change the objective function value. *** correct answer there is no sensitivity output for an integer problem wrong answer MGST391 Illustrative Exams & Solutions Page 14 8. (a) (b) (c) (d) the objective function value (coefficient) *** correct answer the range of values for which a shadow price is valid is false as the sign (+ or -) of the shadow price depends on the direction of the equation (< or >). 9. (d) The solution is 3 since all of the constraints are binding (final value = Constraint R. H. Side or shadow price is not 0) 10. (c) We do not want to make product A. For every item we make above the minimum, the impact on the profit is the shadow price of –$10.00. 11. (a) We must compare the cost and benefit. 50 kg is more than we want (allowable increase), but we would have to take it all. However, we will only use 40 kg with a profit of $8.00 for every extra kg of product so the additional profit would be $320. This is less than the $350 cost. So, the conclusion is not to take it. 12. (d) The shadow price of 9.25 is valid to an increase of 24 units ($222) for an increase of profit to $972 (750 + 222). We cannot tell what the impact is on the profit beyond the allowable increase. 13. (a) (b) Definition is for shadow price, not reduced cost. The solutions are integer but not defined as a constraint. If it was an integer problem, at least one of the changing cells would be identified as integer in the Solver Parameter constraint box. As well, there would not be a Sensitivity report. The optimal solution must also include the optimal profit (value of the target cell). *** correct answer (c) (d) 14. Let X1, X2 and X3 represent binary decision variable where a value of variable Xi equal to 1 represents the performance of project i. Which formulation includes the constraint that one can only perform project 1 or project 2 (but not both) when project 3 is performed? (a) (b) (c) (d) (e) X1 + X2 < X3 Correct Project 1 and 2 can never be performed at the same time X1 + X2 < 2 X3 No because X3 = 1 when X1 = X2 = 1 X1 < X3 and X2 < X3 No because X3 = 1 when X1 = X2 = 1 Two of the above are correct None of the above are correct X1 0 0 1 1 X2 0 1 0 1 X3 0/1 1 1 - MGST391 Illustrative Exams & Solutions Page 15 15. Decision Variables: X = amount of feed X to feed (in kg) Y = amount of feed Y to feed (in kg) Z = amount of feed Z to feed (in kg) MINIMIZE 21 X + 42 Y + 36 Z subject to Carbohydrates (g) 50 X + 20 Y + 30 Z > 60 Protein (g) 15 X + 70 Y + 45 Z > 45 Vitamin A (mg) 10 X + 25 Y + 5 Z > 35 Fat minimum (g) 14 X + 12 Y + 7 Z > 15 Fat maximum (g) 14 X + 12 Y + 7 Z < 35 Max Feed X (kg) Non-negativity X < 0.40 (X+Y+Z) X, Y, Z > 0 Note that this question is easier than the actual question on the exam. MGST391 Illustrative Exams & Solutions Page 16 16. Solution for Mugsy’s Muffins: Note that this question is longer than the actual question on the exam. 17. (a) (b) (c) (d) True – Changing cell C9 has a final value of 400. True – The production used = production available for all three products False – There are 6 changing cells (but 2 are zero in the optimal solution) True – Product 2 has 200 units made at Plant A and 150 units made at Plant B. 18. (b) B has 500 capacity used compared to 650 capacity available 19. (c) The values would go to 0. It is feasible since all of the values are less than or equal to the availability. It would also minimize cost at $0. However, this is not the information that you would be looking for. MGST391 Illustrative Exams & Solutions Page 17 20. (a) 21. MAXIMIZE Subject to calculated from Objective Coefficient – Reduced Cost = 13.00 – 3.00 = 10.00 4N + 3R (Profit, $000) 4N 6N 4N + + + 6R 6R 2R N, R < < < > 48 54 32 0 (grapes, tonnes) (storage space, meters3) (processing time, hours) (non-negativity) Label the constraints and feasible region clearly. Show solution method (iso-profit for $24,000 or corner point). Explicitly state the optimal solution. Co-ordinates (N,R) (N= 0 , R= 0) (N= 0, R=8 ) (N = 3, R = 6) Profit (N= 7, R = 2) 34 0 24 30 RED (N=8, R=0) 32 The optimal solution is to make 7 – 100 litre batches of nectar and 2 – 100 litre batches of red for a maximum profit of $34,000. Optimal solution NECTAR MGST391 Illustrative Exams & Solutions Page 18 22. (b) Forecast = 60 + 55 + 56 + 53 = 56 4 23. (b) Forecast = 60(0.4) + 55(0.3) + 56(0.2) + 53(0.1) = 57 24. (c) ForecastJuly 7 = 64 Ft+1 = α At + (1 – α) Ft where α = 0.2 ForecastJuly 14 = 0.2(60) + 0.8(64) = 63.2 25. Yt 50 65 53 56 55 60 Total Ft 56 56 56 56 56 56 Error = Yt -Ft -6 9 -3 0 -1 4 3 |Error| 6 9 3 0 1 4 23 (Error)2 36 81 9 0 1 16 143 Total 9750 9880 10070 29700 = 3300 9 Growth: 4%*3300 = 132 Bias = sum et = 3 = 0.5 n 6 MAD = sum | et | = 23 = 3.8 n 6 SE = SQRT [sum (et )2 / n ]= SQRT (143/6) = 4.9 26. (c) is referring to causal forecasting. 27. Year 1 2 3 Total Jan-Apr 4010 4070 4050 12130 May-Aug 2670 2590 2610 7870 Sep-Dec 3070 3220 3410 9700 Seasonal Index 4043.3 3300 = 1.23 (3300+132)* 1.23 = 4221 2623.3 3300 = 0.79 3300+2(132)* 0.79 = 2816 3233.3 3300 = 0.98 3300+ 3(132)* 0.98 = 3622 Forecast for Year 4 MGST391 Illustrative Exams & Solutions Page 19 28. (d) a consistent trend pattern implies a steady upward or downward state in the form of a line; the three averaging methods listed will lag behind the trend. 29. (a) single (or simple) exponential smoothing does not involve trend 30. (e) *** all of the above are reasons for incorrect forecasts 31. (b) ***change to a model that includes a seasonality component 32. (c) ***either set the first time period’s forecast equal to the actual demand, just test it with the last few time periods here, or test it against new data MGST391 Illustrative Exams & Solutions Page 20 MGST391 Business Analytics Illustrative Final Exam Questions This set of questions is illustrative of the type of questions to expect on the final exam. As such, this set of questions may not include all the aspects of the course which are covered on the final exam. This practice exam is meant to convey exam style and structure rather than content. 1. Draw a decision tree for Jack’s problem below. At each decision node, clearly label each alternative. Include probabilities at all chance nodes. DO NOT try to SOLVE this decision tree – just draw the tree and label it appropriately. Jack has built a new cottage near the lake and needs water. He is trying to decide if he should connect his cottage to the nearby town’s water supply or drill a well. If he drills (without any test), he expects to be successful 75% of the time. Jack can also hire a firm to do further testing on his property. This would cost him $800. He expects the test results to be favourable about 60% of the time. In either case, he would still need to decide to drill a well or connect to the town. If the test results are favourable, he expects his drilling to be successful 80% of the time. If the test results are unfavourable, he expects the drilling results to be unsuccessful about 90% of the time. MGST391 Illustrative Exams & Solutions Page 21 2. A retailer is trying to determine whether to order 1, 2, or 3 magazines. The payoff table for profit is: Demand (# of magazines) Ordering Decision 1 2 3 1 2 -2 -3 2 -1 4 0 3 -3.5 2 6 The probabilities of a demand of 1, 2, or 3 magazines is 0.4, 0.25 and 0.35 respectively. What is the retailer’s best decision, based on the MAXIMAX, MAXIMIN, MINIMAX Regret, and EMV criterion? MGST391 Illustrative Exams & Solutions Page 22 3. Decision trees are particularly useful when: (a) (b) (c) (d) 4. The expected value of sample information (EVSI) can be used to (a) (b) (c) (d) 5. establish a maximum amount to spend on getting information calculate conditional probabilities establish risk avoidance none of the above An insurance company plans to market a particular type of term life insurance. The premium that will be paid for this depends on the expected amount that the insurance company will have to pay out on each policy plus any cost and profits. Assume the company wishes to make $50 profit on each policy and there is $40 overhead cost. The probability a person in a certain age group would die in the coming year is 0.001. How much would the company charge for a $100,000 life insurance policy? (a) (b) (c) (d) (e) 6. perfect information is available formulating a conditional values table a sequence of decisions must be made none of the above 90 100 190 999 none of the above A company is trying to determine if they should open a new café or expand their current café. Their situation is shown in the decision tree below. What is the EVPI? (a) (b) (c) (d) 12,900 23,000 120,000 180,000 MGST391 Illustrative Exams & Solutions Page 23 Complete the Bayesian probability tree for this situation (10 values) and answer the four questions that follow. The marketing manager of a toy manufacturing company is considering marketing a new toy. In the past, 35% of the toys introduced by this company have been successful and 65% have been a failure. Before the toy is marketed, market research is conducted and a report, either favourable or unfavourable, is compiled. In the past, 80% of the successful toys received a favourable market research report, and 30% of the failed toys received a favourable market research report. 13. What is the probability of a toy being successful or unsuccessful? a. b. c. d. e. 14. 0.00 0.20 0.40 0.60 0.80 < < < < < p p p p p < 0.20 < 0.40 < 0.60 < 0.80 < 1.00 What is the probability that the report is unfavourable given that the toy was unsuccessful? a. b. c. d. e. 0.00 0.20 0.40 0.60 0.80 < < < < < p p p p p < < < < < 0.20 0.40 0.60 0.80 1.00 MGST391 Illustrative Exams & Solutions Page 26 15. What is the probability that a toy will be successful and it receives a favourable report? a. b. c. d. e. 16. p p p p p < 0.20 < 0.40 < 0.60 < 0.80 < 1.00 0.00 0.20 0.40 0.60 0.80 < < < < < p p p p p < < < < < 0.20 0.40 0.60 0.80 1.00 Which of the following is true? (a) (b) (c) (d) 18. < < < < < What is the probability that a toy will be unsuccessful and it receives an unfavourable report? a. b. c. d. e. 17. 0.00 0.20 0.40 0.60 0.80 In simulation, we always assume that the arrival rates are normally distributed. For randomly generated numbers like service time in a line-up simulation, it is not possible for a randomly chosen value to immediately follow itself. For example, a "5" minute service time could not follow another "5" minute service time. The goal of simulation is to create an environment in which it is possible to obtain information about alternative actions via experimentation. Using simulation is typically less work than using Queuing Theory When simulating demand using @RISK, the average simulated demand over a large number of iterations would be an estimate of the: (a) (b) (c) (d) expected demand maximum possible demand sampled demand daily demand MGST391 Illustrative Exams & Solutions Page 27 19. Dave’s Candies is a small family-owned business that offers gourmet chocolates and ice cream fountain service. For special occasions, such as Valentine’s Day, the store must place orders for special packaging boxes of chocolates several weeks in advance from their supplier. One product, the Valentine’s Day Chocolate Delight, is bought for $7.50 per box and sells for $12.00. Any boxes not sold by February 14 are discounted by 50 percent and can always be sold easily. Historically, Dave’s Candies sold has followed a normal distribution of 65 boxes with a standard deviation of 10 boxes each year. Dave’s dilemma is deciding how many boxes to order for Valentine’s Day customers. If demand exceeds the purchase quantity, then Dave loses profit opportunity. On the other hand, if too many boxes are purchased, he will lose money by discounting them below cost. Fill in the chart above. Identify the following cells: Distribution: ________ Parameter: ______ Output: ______ 20. What are the main purpose of the Fixed Seed Value in @RISK? ___________________________________________________________________ 21. What is the purpose of confidence intervals? (a) (b) (c) (d) To make you do more work (!) To increase your confidence in the answer To give you a range in which the population mean lies To assign a confidence value to a result MGST391 Illustrative Exams & Solutions Page 28 Some simulation output is shown below. Use it to answer the next 2 questions. 22. What is the 95% confidence interval for the Cost for the above example to 4 decimal places? Note that this data is in 00s. Lower 95% C.I. ______________ 23. Upper 95% C.I. _______________ What is the 90% confidence interval for the probability that the Cost will be less than or equal to $13 (in % to 2 decimal places)? Lower 90% C.I. ______________% MGST391 Illustrative Exams & Solutions Upper 90% C.I. ______________ % Page 29 24. Which of the following is TRUE regarding empirical distributions? (a) (b) (c) (d) They are ‘named’ distributions (e.g., normal, exponential) They systematically overestimate real variability They convert observed data into a probability distribution None of the above are true 25. Which of the following is a disadvantage of using simulation? (a) they allow for the study of "what-if" questions (b) each simulation requires a unique model (c) allows the study of interaction of components or variables to determine which are important 26. Which queuing model best represents getting service from a cafeteria? (a) (b) (c) (a) 27. What is the fundamental reason line-ups occur? (a) (b) (c) (d) 28. single-server, single-phase single-server, multi-phase multi-server, single-phase multi-server, multi-phase people are not distracted too few skilled servers temporary mismatch between supply and demand because people don’t complain Jones, the mechanic at Golden Muffler Shop, is able to install new mufflers at an average rate of three per hour (or about one every 20 minutes), according to a negative exponential distribution. Customers seeking this service arrive at the shop on the average of two per hour, following a Poisson distribution. The customers are served on a first-in, first-out bases and come from a very large population of possible buyers. The owner of Golden Muffler Shop, estimates that the cost of customer waiting time, in terms of customer dissatisfaction and lost goodwill, is $10 per hour spent waiting in line. a) What are the values of λ and µ per hour? λ = µ = b) On average, how many cars are in the system? MGST391 Illustrative Exams & Solutions Page 30 c) What is the average waiting time in the system? d) On average, how many cars are waiting in line? e) What is the average waiting time per car? f) What is the utilization rate of Jones? g) What is the probability that there at no cars at the shop? h) How many cars are serviced in a typical 8 hour working day? i) What is the average waiting-time cost for Golden Muffler Shop? MGST391 Illustrative Exams & Solutions Page 31 29. Consider a single-server line-up in a store. A customer will not wait for more than 10 minutes in the line-up. If the service start time is more than 10 minutes after they arrive, cross them off and have the next person in line be served. Assume no customers are in the store at 9:00. Complete the chart below. Customer Interarrival Time (minutes) Arrival Time (hh:mm) Leave? (Yes/No) Service Time (minutes) Service Start Time (hh:mm) Annelies 4 9:04 No 8 9:04 Benson 11 9:15 Carolyn 2 9 David 5 14 Erin 8 11 Fay 7 17 MGST391 Illustrative Exams & Solutions Service End Time (hh:mm) Waiting Time (if they do not leave) 16 Page 32 30. Solve a DecisionTree Kim Fleming has a race horse named Cocoa Futures; she is trying to decide whether to have the horse race for another season or not. If the horse does not race, it will be sold immediately. Another choice for Kim is to send Cocoa Futures to a new trainer who would evaluate the horse at a cost of $5,000 and then make a recommendation as to whether or not to continue racing. (Note that this cost IS included on the end nodes of the decision tree.) Solve the decision tree below. Ensure to prune the tree and to highlight the optimal strategy. a) What is the proper course of action for Kim to take? b) What is the EMV of the optimal solution? c) What decision should Kim make if the horse was not evaluated by a trainer? d) What is the expected profit if the trainer evaluation is not available? e) If the horse was checked by a trainer and the trainer recommended ‘do not continue’ but Kim decided to race for the season, what is the expected profit? f) What is the expected value of sample information (EVSI)? g) What is the expected value of perfect information (EVPI)? MGST391 Illustrative Exams & Solutions Page 33 $1,245,000 MGST391 Illustrative Exams & Solutions Page 34 MGST391 Illustrative Final Exam SOLUTIONS 1. MGST391 Illustrative Exams & Solutions Page 35 2. The Payoff Tables: MGST391 Illustrative Exams & Solutions Page 36 3. (c) a sequence of decisions must be made 4. (a) comparing the value to the cost of information 5. (c) The expected payout per policy is 0.001x100000 = $100. The overhead plus profit is $90. So they need to charge $190. 6. (a) MGST391 Illustrative Exams & Solutions Page 37 Solved Decision Tree for Questions 7 to 11: 7. (a) EMV for Large = -100, Small = -35, None = -20 8. (b) EMV for Large = 30.4, Small = 26.4, None = 0 9. (a) EMV = 40.6 10. (e) The course of action has to specify entire sequence. a), b) and c) don’t do that. d) is wrong because if the study is negative, doing nothing (EMV = -20) is better than building a large plant (EMV = -100) or small plant (EMV = -35). If the study is positive the best option is a large plant with an EMV = 101.6 11. (a) EMV = -20 12. (d) EVSI assumes sample information is free, i.e., we have to add back the survey cost to EV with SI. So EVSI = $63,000 – 50,000 = $13,000 MGST391 Illustrative Exams & Solutions Page 38 See diagram at bottom of page. 13. 1.00 (all of the possibilities) 14. 0.70 15. 0.28 16. 0.455 17. (c) The goal of simulation is to create an environment in which it is possible to obtain information about alternative actions via experimentation. 18. (a) expected demand [mean or average value] 0.8 * 0.35 = 0.28 0.2 * 0.35 = 0.07 MGST391 Illustrative Exams & Solutions Page 39 19. Distribution Cell: B3 Parameter Cell: B6 Output Cell: C11 20. The purposes of the Seed value in @RISK is: 1) Specifying a non-zero RNG seed @RISK causes repeat runs to use the same random numbers, which makes the results of multiple model runs more comparable. 21. (c) 22. CI for mean: sample mean + z-value * mean standard error where MSE = sample standard deviation / square root of the number of observation Lower 95% CI: 11.6998 – 1.96 * (1.5494 / sqrt (1000)) = 11.6038 Upper 95% C.I.: 11.6998 + 1.96 * (1.5494 / sqrt (1000)) = 11.7958 23. CI for proportion is p-bar + z * sqrt(p-bar * (1- p-bar)/n Lower 90% C.I.: 0.778 - 1.645 * SQRT (0.778(0.222)/1000) = 0.7564 = 75.64% Upper 90% C.I.: 0.778 + 1.645 * SQRT (0.778(0.222)/1000) = 0.7996 = 79.96% 24. (a) (b) (c) Should be theoretical Should be underestimate *** Correct answer 25. (b) *** Correct answer 26. (b) one server per station, many stations 27. (c) variability in arrivals and service rates To give you a range in which the population mean lies *** within a certain confidence level MGST391 Illustrative Exams & Solutions Page 40 28. MGST391 Illustrative Exams & Solutions Page 41 29. Customer Interarrival Time (minutes) Arrival Time (hh:mm) Leave? (Yes/No) Service Time (minutes) Service Start Time (hh:mm) Service End Time (hh:mm) Waiting Time (if they do not leave) Annelies 4 9:04 No 8 9:04 9:12 0 Benson 11 9:15 No 16 9:15 9:31 0 Carolyn 2 9:17 Yes 9 David 5 9:22 No 14 9:31 9:45 9 Erin 8 9:30 Yes 11 Fay 7 9:37 No 17 9:45 10:02 8 MGST391 Illustrative Exams & Solutions Page 42 $930,000 MGST391 Illustrative Exams & Solutions Page 43 $1,245,000 MGST391 Illustrative Exams & Solutions Page 44