: Chapter Three Forecasting 107 a associative model, 73, ‘mean squared error (MSE), 95 » KEY TERMS : bias, 98 moving average, 77 centered moving average, 88 naive forecast, 75, control chart, 97 predictor variables, 90 correlation, 93 random variations, 75 cycles, 74 regression, 90 Delphi method, 74 ‘seasonal relative, 86 4 error, 95 ‘seasonal variations, 85 exponential smoothing, 19 ‘seasonality, 74 forecast, 69 time series, 74 irregular variation, 74 time-series forecasts, 73, judgmental forecasts, 73 ‘tacking signal, 98 Feast squares lie, 90, ‘end, 74 linear trend equation, 81 ‘nend-adjusted exponential emoothing, £4 mean absolute deviation (MAD), 95 ‘weighted average, 79 mean absolute percent error (MAPB), 95, a i L | Forecasts based on averages. Given the following dain: Problem 1 : Nani of a esi Compas ; 1 4 2 6 3 ps : 4 as : 5 Propere a forecast using each of these approces: 4 The appropriate nuve approach A treeperod moving verge. © | «A weighted average using weights of $0 (most recent), 30, and 20 4. Bxponenisl smoothing with a smoothing constant of 40. | 2 Te values ate stable (., no tendo eels). Therefore, the most cent value ofthe ses Solution : becomes the net forcast 6 © | nay BABE Ay : 3 ce, Fe 20655) + 30(58) + 50(64) = 604 : Pied Caleta 1 The previous vate ofthe sre used 2 2 to srt oacos] 3 60 + 40(5 ~ eo) ~ 62 : a 62+ a5 ~ 2) = 532 . 5 58.2 + 40(68 - 69.2) = £8.72 . 6 sera + ae ~ 527) = ena Be—=4 108 Part Two Forecasting ‘You also can obtain the forecasts and a plot using an Excel template, as shown: Xcel | | | | | i | ii i yi ra sili fall AER id Problem 2 Using seasonal relatives. Apple's Citrus Fruit Farm ships boxed fruit anywhere in the wotld, Us Ht the following informatio, a manager wants to forecast shipment forthe fet fou months of ty vyeat. aly Seasonal | Mont } van, - Whee OB | Fea Fug cicee 8 qd Mec sep 7 Ry] Ey May ae Now sot u E | Jun. o7 Dec. Ww ‘The monthly forecast equation being used is | | | ty Fa 402+ 30 i] where b fo = January of /ast year | 2, = Nomber of sipments tid Solution 4% Deterine trend amounts forthe frat four months of next year January, ¢ = 245 Fobra { t= 25; ete, Thus, yey | Fray = 402 + 3(25) = 477 ' ' Fre = 402 + 3(26) = 480 i {| Frage = 402 + 3(27) = 483 i ', Multiply each monthly tend by the coresponding seasonal relative for that month, it | Seasonal 1] | ela j Pid un, 12 si) = 008 Pda Feb 13 aniiai= ens iid Mer) aa 4ea(.3} = e240 | Apr 4 se.) 19 |Chapter Three Forecasting 109 Linear trend line. Pot the data on a graph, and verify visually that linear tend line is appro- priate. Develop a line trend equation forthe following data. Then use the equation to predict the next two values ofthe series. Pete 4 8 ” se 8 8 0 % ee ‘A plot of the data indicates that a linear tend line is appropriate: Is nae Demand, y 9 4“ From Table 21, wth n= 8 2 08 m 160 t= AS and Bet = 286 5 26 5 25 5 330 a a0 B 4 2 e 2; 15(488) (a iscasy 175 pw BILBEE $88 = 11505) sy 9 ‘Thus, the trend equation is F; = 45.47 + 1.75¢. The next two forecasts ate: Fro = 4847 + 1.75(10 Fy = ASAT + 1.75(11) = 64.12 97 Problem 3 Solution112 Part Two. Foceastng Xcel ‘The graph supports & linear relationship, Ca 'b. You can obtain the regression coefficients using the appropriate Excel template. Simply replac the existing data for x and y with your deta. Note: Be careful to entec the values for the vari able you want to predict as y values. In this problem, the objective is to prediet sales, #0 th soles values are entered in the y cohumn, The equation is yz = 7.129 = 4.275x, ©, Forx=5, y.= 7129 + 4.275(5) = 28.50. Problerh 7 Accuracy and contol of forecasts. The menager ofa age manufacturer of industri! pumps mus / choose Between tno altematve forecasting techniques. Both techniques have been used to prepare Foyecasts for sitsmonh period. Usieg MAD af a eiterion, which tchnige hat the beter pet Farmance reson? FoREcAsT Month Demand mique 1_Technigee 2 Vo @ a % a aw @ Be 6 0 ” fe so @ so ae "7 py fo ae 0 2 Solution Check tt ech forecast san verge ctor of approximately zzo (See computations that follow) Month 0 1 2 a oo 5 = 3 2 ™ ee -“ ow we “2a 3 8 @ rics 7 7 ‘ we 19 aa oo wos 5 we 7 aE 2 6 6 6 ps © fey 2 Ereteea 2 OH MAD, Ziel Biel MAD, elChapter Three Forecasting 113 ‘Technique 1 is superior in this comparison because its MAD is smaller, although six observations would generally be too few on which to base a realistic comparison. Control chart. Given th demand data tht follow, prepare a nve forecast for period 2 through 10. Problem 8 “Then determine cach forecast ero, and use those vals to oblain 2s contol limits. If demand in {he next tw periods tums out tobe 125 and 130, ean you conciade that the forecasts are in contol? Petdes--ee a Heeetee aeeeen eee eee ete tep eeeeg te eC Danand: 87mm For a naive forecast, each period's demand becomes the forecast for the next period. Henee, the Solution forecasts and errors are: Peod Demand ___Forocast__Eror__Enor™ 1 18 = - = 2 ™ wo 4 1 3 va nr 3 8 4 nia m4 1 5 1 ng 7 48 8 ry m4 8 7 uw mo a 1 u 5 8 ° a m2 ‘ 0 ra m3 8 7 ito [Senor [AS 433 (a = Numberoferors) a= V9=1 “Tho contol limits are 2(4.38) = £8.65 “The forecast fr period 11 was 124. Demand tured otto be 125, for an etrrof 125 ~ 124 = +1 “This is within the limite of 28.66. I the next demand is 130 andthe nave forecast is 125 (base on the pericd 11 demand of 125), the exo is +5. Again, this is within the limits, s you eannot con- clude the forrat i not wocking propery. With more values—at leas ve or six—you could plot the csvus to see whether you could detect azy pattems suggesting the presence of onrandomness 1. What are the main advaneages that quaniative techniques for forecasting have over qualitative DISCUSSION techniques? What inittions do quantitative techniques have? AND REVIEW 2 What are some of the consequences of poor forecasts? Explain QUESTIONS List the specific weaknesses of each of these approaches to developing a forecast: a. Consumer surveys. , Salesforce composite ©. Committee of managers or executives 4, Briefly describe the Delphi technique. What are its main benefits and weaknesses? 5, What is the purpose of establishing control limits for forecast errors? 6, What factors would you consider in deciding whether to use wide or narrow control limits for forecast? 7. Contrast the use of MAD and MSE in evaluating forecasts, 48, What advantages as a forecasting tool does exponential smoothing have over moving averages? 9. How does the number of periods in a moving average affect the responsiveness of the forecast? 10, What factors enter into the choice of a value forthe smoothing constant in exponential smoothing? 11, How accurate is yeur local five-day weather forecast? Support your answer with actual data, 12, Explain how uting a centered moving average with a length equal to the length of a season. eliminates seasonality from a time series.i +4. Chapter Three Forecasting 1, Prodict orders forthe following day foreach of the products using an appropriate naive method. Hint: Plot each dataset 'b. What should the use of sales data instead of demand imply? ‘National Scan, Inc, sells radio frequency inventory tags. Monthly sales fora seven-month period were as follows: Sales ‘Month {000 units) Feb, 19 Mac. 8 Ape - 8 Moy » Jun sees 6 ul 2 Ag. » ‘a Plot the monthly data on a sheet of grapia papet ‘Forecast September sales volume using each of the following (1) A linear trend equation @) A five-month moving average. G) Exponential smoothing with a smoothing constant equal to .20, assuming a March fore- cast of 19(000).. (@) The naive approach. (5) A weighted average using .60 for August, .30 for July, and .10 for June ce. Which method seems least appropriate? Why? (Hint: Refer to your plot from part a.) 4 What does use of the erm sales rather than demand presume? A dry cleaner uses exponential smoothing to forecast equipment usage at its main plant, August usage was forecast 10 be 88 percent of capacity; actual usage was 89.6 percent of capacity. A smoothing constant of .1 is used. a. Prepare a forecast for September. ', Assuming actual September usage of 92 percent, prepare a forecast for October usage ‘An electrical contractor's records during the Inst five weeks indicate the number of job requests: Weske 1 2 3 4 5 fewest; = Prodiet the number of requests for week 6 using each of these methods: a. Naive. . A feur-period moving average. «. Exponential smoothing with « = .30. Use 20 for week 2 forecast. ‘A cosmeties manufactures marketing department has developed a linear tend equation that ean ‘be used to predict ennual sales of its popolar Hand & Foot Cresm. F,= 80+ 150 where F, = Annual sales (000 bottles) 15 0 corresponds to 1990 1° Are annual sales increasing or decreasing? By how much? Predict annual sales for the year 2006 using the equation From the following graph, dorermine the equation of the linear trend line for timeshare sales for Glib Marketing, Ine 115116 Part Two Forecasting Annual Sales, Gib Sales, Inc 600 ‘500 0 200 200 | 7. Freight car loadings over a 12-year period at a busy port are as follows: Sales (units) Determine a linéar trend line for freight car loadings. Use the trend equation to predict loadings for weeks 20 and 2 | ‘The manager intends to install new equipment when the volume exceeds 800 load | ‘week. Assuming the current trend continues, the loading volume will reach that level in imately what week? | 1a fj ‘Week Num Week Number Number iy Hl oe ee ti 2 6 8 30 “ 475 al 4 8 0 we 6. 510 it 5 0 n ~ ” 5 8, a, Obtain the linear trend equation for the following data on new checking accounts Savings Benk and use it to predict new checking accounts for periods 16 through 19. | Now New Now nn Period counts Period Accounts Accou | 1 00 6 me " ea | 2 m4 T 28 Ri TB | 3 m 8. zo 8 0 | 4 ne 2 ES Wi 8 | 5 28 0... a 6 10 b, Use trend-adjusted smoothing with 4, What is the forecast for period 16? 9. After plotting demand far four periods, an emergency room manager has concluded that adjusted exponential smoothing model is appropriate to predict furure demand. The init ‘mate of tend is based on the net change of 30 for the diree periods from 1 to 4, for an of +10 units, Use = 5 and B = .4, and TAF of 250 for period 5. Obtain forecasts for 3 and B = 2 to smooth the new account date Bose 1 i Git est =r =o 4 1 0 285 | ne nm 1 3 28 25 oe a i }