Review problem 1 Text problem 9

Review problem #1 Text problem 9 - 6, 7 • Weekly sales (in hundreds) of PERT shampoo at the Save. Mor drug chain for the past 16 weeks are as follows: Management wishes to forecast the annual demand for PERT shampoo in order to determine an optimal order policy. • Graph this time series. A stationary model seems to properly describe this time series. Go to file ch 9 -6. xls and create the graph. 1

• Verify statistically that a stationary model is appropriate forecasting this time series. Do not reject the null hypothesis. There is H 0: b 1= 0 insufficient evidence for the existence H 1: b 1¹ 0 of a trend at (even) 15% significance level. 2

• Compare the accuracy of the following stationary models: – 3, and 6 period moving average – 3 period weighted moving average with optimal weights vs. Exponential smoothing with optimal alpha 3

Comparing moving average models 4

Obtaining the optimal weights 3 - period moving average The sum of the weights appears in cell c 9. Don’t forget to select “Assume non-negative” in “Options” 5

The optimal alpha Exponential smoothing model Don’t forget to select “Assume non-negative” in “Options” 6

Review problem #3 A firm that manages and maintains several apartment house complexes has experienced the following expenses over the last five years. Quarter : t 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Yt 27 39 41 50 44 47 58 63 73 92 94 102 123 119 128 147 155 158 164 187 (a) Plot this time series using an Excel graphical tool (use an x-y scatter plot) (b) Run a linear regression model and compare its performance to the best Holt’s model you can obtain, when optimizing alpha and gamma. Use the MAD criterion to compare. (c) Repeat part (b), this time optimize alpha , gamma, the initial forecast and the initial trend. Which model do you select now based on the MAD? 7

Review problem 3 - Solution • The plot indicates the presence of a linear trend. • Thus, a linear regression model or the Holt’s model should perform well. 8

Review problem 3 - Solution • Running linear regression and checking the pvalue of the t-test for b 1 confirms the observation. The time series has a linear trend. The p-value is very small, implying that there is sufficient evidence to support the hypothesis that b 1¹ 0 9

Review problem 3 – Solution The Linear regression Forecasting model 10

Review problem 3 – Solution Optimizing alpha and gamma For Holt’s Alpha and gamma are both £ 1 Notice: Only alpha and gamma are optimized Don’t forget to select “Assume non-negative” in “Options”, since alpha and gamma are both 11 ³ 0

Review problem 3 - Solution Optimizing alpha, gamma, L 0, T 0 for Holt’s Notice: Both alpha, gamma, L 0 and T 0 are optimized Don’t select “Assume non negative” in “Options”, since L 0 and T 0 might be negative. But then, add the non-negativity constraints of alpha and 12 gamma

Review problem 3 – Solution • The Holt’s model with all the parameters optimized had the best performance over the past 20 periods based on MAD. – Holt’s (all parameters optimized) MAD = 5. 39. – Holt’s (only a and g are optimized) MAD = 6. 89 – Linear regression MAD = 7. 17 13

Review problem 4 • • Analyze the following time series. Compare the following four forecasting techniques, and determine which one you would use to make a forecast. (i) The exponential smoothing (ii) The weighted moving average (iii) The Holt’s method (v) The linear regression model. Write down each optimized model. Perform a forecast by hand for period 61 and 62 using the template results. Show you would calculate MAD for the first 4 periods of each technique. 14

Review problem 5 • Sport King is a chain of 12 sporting goods shops in South Carolina. Total chain demand for the Whamco flying disk averages 250 units a week at a cost. While the disks come in various colors, the firm’s policy is to determine the appropriate total order quantity using the EOQ formula and divide this total among the various colors in proportion to historic demand. • Sports King buys the Whamco flying disks for $2. 50 each, and sell them for $3. 50 each. Sport King uses an annual inventory holding cost rate of 28%, and the cost to place an order is $40. Lead-time for delivery is two weeks, and Sport King desires a safety stock of 50 units. 15

Problem 5 - continued • Determine: – – The optimal order quantity of disks. The number of calendar days between orders (cycle time). The reorder point for the disks. The total annual inventory cost (holding, ordering, procurement) of this policy. • Solution • Annual demand = 250*52 = 13000 disks • Unit holding cost =. 28(2. 5) = $. 7 per unit per year – The optimal order quantity 16

Problem 5 - continued – The cycle time T = Q*/D • T = Q*/D = 1218. 89/13000 =. 09376 years = (. 09376)(52) = 4. 875 weeks = 4. 875*7 = 34. 129 calendar days. – The reorder point R = inventory position = Inventory at hand + Outstanding orders. • There are no outstanding orders in our problem: Reorder point 2 weeks 4. 875 weeks • Reorder point = Inventory at hand = (Lead time in weeks)(demand per week) = 2*250 = 500 17

Problem 5 - continued – Total annual inventory costs = Ann. Ordering cost Ann. Holding cost Procurement cost 18

Problem 5 - continued – Whamco offers its customers the following all-units price discount schedule: Number Ordered Price per Unit 1– 499$ 500– 999 1000– 1999 2000– 4999 5000– 9999 10000 and above 2. 50 $2. 25 $2. 10 $1. 90 $1, 80 $1. 75 19

Problem 5 - continued – Assuming that holding costs are discounted, determine the following: a. The optimal order quantity of Whamco disks. b. The reorder point for the disks. c. The number of calendar days between orders (cycle time). d. The total annual inventory cost (holding, ordering, procurement) for this policy. 20

Review Problem 6 • The Compland Computer Store sells the Hew. Pac. X laser jet printer. Demand has been averaging 24 units per six-day week. The printers cost Compland $1095 each, and Compland uses a 20% annual inventory holding cost rate. The cost of placing an order with Hew. Pac is $90, and the lead time is eight working days. • If Compland runs out of the Hew. Pac. X printer, it gives its customers a loaner printer until the next shipment of Hew. Pac. X printers arrive. Each loaner costs Compland $1. 50 per calendar day. • Determine: – – The optimal order quantity of Hew. Pac. X printers. The reorder point for these printers. The proportion of customers who will get loaner printers. The total annual inventory cost (holding, ordering, shortage, and procurement) of this policy. 21

Problem 6 - continued • Solution D = 24*52 = 1248 per year. Ch = IC =. 20(1095) = 219 per unit per year. Co = 90. Cs = 1. 50(365) = 547. 5 per unit per year. Cb = 0 (not included) – The optimal order quantity = – The reorder point R = LD-S*. The optimal backorder is Then R = 8(24/6) – 10. 82 = 21. 18 22

Problem 6 - continued – Proportion of customers on backorder (assuming each customer orders one unit) = S*/Q* = 28. 6% – Total inventory costs = Q = 37. 89 27. 07 27. 89 R = 21 8 days Lead time Cycle -10. 82 23

Review Problem 7 • Return to the Sports King disk sale problem. – Management has decided to consider uncertainty in the demand pattern. A statistical analysis revealed that the weekly demand is normally distributed with: • • mean = 250 disks per week Standard deviation = 80 disks per week. Lead time is 2 weeks Required cycle service level is 92%. – Find • The order quantity, the reorder point, the safety stock • The total inventory costs 24

Problem 7 - continued • Solution – The order quantity does not change because we assume an EOQ model, as before. – The reorder point depends on the service level. P(DL > R) =. 08 P(Z > Z. 08) =. 08 but Z. 08 = 1. 405 so (R – m)/s = 1. 405 or R = m L + Z. 08 s. L. – Before making the calculation, we need to adjust the mean and standard deviation of the demand during the lead time. m. L = 2(250) = 500; s. L 2 = (2)(80)2 = 12800 s. L = 113. 13 2 weeks variance per week – Calculating the reorder point: R = 500 + 1. 405(113. 13) = 658. 94 – The safety stock = R – m. L = 1. 45(113. 13) = 164. 03 disks. 25

Review problem 1 Text problem 9 — 6