Scribd Logo
Upload

Welcome to Scribd!

  • 152 views

    HW4

    Uploaded by

    timmyneutron
    1. This homework assignment involves forecasting sales data using various techniques like seasonal factors, linear regression, moving averages, and exponential smoothing. 2. Students are asked to analyze time series data on sales of tennis shoes, theme park arrivals, and basketballs to develop forecasts and compare forecasting methods. 3. Additional problems involve using Holt's exponential smoothing to forecast stock prices, chocolate production, and chemical shipment handling times. Students are to select optimal parameters and evaluate forecast accuracy.

    Copyright:

    Attribution Non-Commercial (BY-NC)
    Download as PDF, TXT or read online from Scribd

    HW4

    Uploaded by

    timmyneutron
    152 views3 pages
    1. This homework assignment involves forecasting sales data using various techniques like seasonal factors, linear regression, moving averages, and exponential smoothing. 2. Students are asked to analyze time series data on sales of tennis shoes, theme park arrivals, and basketballs to develop forecasts and compare forecasting methods. 3. Additional problems involve using Holt's exponential smoothing to forecast stock prices, chocolate production, and chemical shipment handling times. Students are to select optimal parameters and evaluate forecast accuracy.

    Copyright

    © Attribution Non-Commercial (BY-NC)

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    1. This homework assignment involves forecasting sales data using various techniques like seasonal factors, linear regression, moving averages, and exponential smoothing. 2. Students are asked to analyze time series data on sales of tennis shoes, theme park arrivals, and basketballs to develop forecasts and compare forecasting methods. 3. Additional problems involve using Holt's exponential smoothing to forecast stock prices, chocolate production, and chemical shipment handling times. Students are to select optimal parameters and evaluate forecast accuracy.

    Copyright:

    Attribution Non-Commercial (BY-NC)
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    152 views3 pages

    HW4

    Uploaded by

    timmyneutron
    1. This homework assignment involves forecasting sales data using various techniques like seasonal factors, linear regression, moving averages, and exponential smoothing. 2. Students are asked to analyze time series data on sales of tennis shoes, theme park arrivals, and basketballs to develop forecasts and compare forecasting methods. 3. Additional problems involve using Holt's exponential smoothing to forecast stock prices, chocolate production, and chemical shipment handling times. Students are to select optimal parameters and evaluate forecast accuracy.

    Copyright:

    Attribution Non-Commercial (BY-NC)
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    of 3

    ISyE 3013 Supply Chain Modeling: Logistics

    Spring 2008
    Homework 4
    Due: February 21, 2008

    100 points.

    This assignment is to be done individually - you must do (and show) your own
    work.

    Part I. This portion of the homework assignment is to be completed by


    hand and submitted in class on the due date. You must show your work
    to receive full credit.

    1. The amount of tennis shoes that Toe Locker sells quarterly exhibits a constant mean
    over time, but sales fluctuate depending on the quarter of the year. Develop a set of
    seasonal factors that could be used to forecast tennis shoe sales using three years worth
    of quarterly data that is provided.

    Quarter Year 1 Year 2 Year 3


    1 12 16 14
    2 25 32 45
    3 76 71 84
    4 52 62 47

    2. Using average daily arrival rates for the months of January through June 2007, a
    Magic World theme park logistics engineer fitted a simple linear regression model
    to the data. The single causal factor considered was a linear trend. The resulting
    regression coefficient estimates were: Â0 = 1087.4 and Â1 = 230.3143 where Â0 is the
    estimate for the level (or intercept) and Â1 is the estimate for the trend (or slope).
    Use this model to predict average daily arrival rates for July through December 2007.

    3. Sales of basketballs at Let’s Play Ball Sports exhibit a constant mean. The demand
    for March through December 2007 was 90, 95, 86, 112, 93, 95, 88, 91, 115, 84.

    (a) Using moving average with m = 3, create the one-step ahead forecasts for June
    through December 2007. Compute the errors and the MAD.
    (b) Using exponential smoothing with α = 0.2, create the one-step ahead forecasts
    for June through December 2007. You will still initialize your method in March
    (as shown in class) but only compute the errors and MAD using the months June
    through December. This will keep your “start-up” forecasts from affecting the
    MAD and make comparison between the two methods meaningful.
    ISyE 3103 · Spring 2008 · Homework 4 2

    (c) Based on MAD, which method performed better?


    (d) Using your moving average model, what is the multiple step ahead forecast for
    March 2008 created in December 2007?
    (e) Using your exponential smoothing model, what is the multiple step ahead forecast
    for March 2008 created in December 2007?

    4. Weekly sales for brown eyeliner from Paint Your Face exhibit a linear trend. The
    demand for weeks 1 through 10 was 96, 99, 100, 102, 98, 106, 112, 121, 118, and 126.

    (a) Calculate the forecasts for weeks 2 through 10 using Holt’s double exponential
    smoothing with α = 0.1 and β = 0.15.
    (b) What is the multiple-step ahead forecast for period 12 created at the end of period
    10?
    (c) What is the multiple-step ahead forecast for period 12 created at the end of period
    6?

    Part II. This portion of the homework assignment is to be completed in Ex-


    cel and submitted via email to myself and Viktoriya by class time on the due
    date. The data you will need to complete these problems can be found in
    the spreadsheet that was attached to the email when the homework was sent
    out. When you email your file to us, please use ISyE3103 HW4 in the sub-
    ject line of the email, and attach your file as Firstname.Lastname.HW4.xls.

    5. The data set “google” gives the closing stock prices of Google over a time period of
    252 trading days. For investment purposes, you want to forecast the closing stock
    price the next day. Forecast the closing stock price for the next day using each of the
    following forecasting methods. You should do this for as many (historical) periods as
    possible and calculate the errors. For each method, experiment with different model
    parameters to find the best-fit model of each type. The file you turn in should present
    your best-fit model of each type (your best-fit moving average model, best-fit simple
    exponential smoothing model, and best-fit double exponential smoothing model). For
    each of these three best-fit models, you should also plot the errors versus time and
    calculate the mean square error. Based on this analysis and any other analysis you
    think is important, compare the quality of the three forecasting methods on this data
    set.

    (a) Moving average: parameter m


    (b) Simple exponential smoothing: parameter α
    (c) Double exponential smoothing: parameters α and β

    6. The data set “chocolate” gives the historic monthly production for chocolate (in tons)
    in Australia from 1985 to 1994. For capacity planning purposes you want to build
    ISyE 3103 · Spring 2008 · Homework 4 3

    a forecasting model for the Australian chocolate industry. The data exhibits both
    trend and seasonality. Forecast the production levels for as many (historical) periods
    as possible using Holt’s double exponential smoothing with constant multiplicative
    seasonal factors. Calculate the errors, plot the errors versus time, and calculate the
    mean squared error. Experiment with different factors for α and β and present your
    best results (we only need to see your best-fit model and error analysis).

    7. The following data were collected from shipments of chemicals in drums arriving at a
    warehouse in the last 20 days (Monday through Friday for the last 4 weeks). Variable
    f1 denotes the number of drums in the shipment, variable f2 denotes the total weight
    of the shipment in hundredweight, and X denotes the number of minutes required to
    handle the shipment each day.

    t 1 2 3 4 5 6 7 8 9 10
    f1t 5 6 12 17 6 7 8 15 14 10
    f2t 11.2 10.2 25.0 29.8 11.3 13.1 19.4 28.3 32.7 25.3
    Xt 62 66 121 154 77 72 100 132 141 129

    t 11 12 13 14 15 16 17 18 19 20
    f1t 5 7 16 20 11 9 14 16 21 23
    f2t 12.6 13.9 34.8 29.1 24.1 19.4 25.8 37.8 44.3 37.3
    Xt 80 93 143 177 141 101 152 169 183 152

    (a) Consider forecasting the daily time required to handle shipments of chemical
    drums, by developing a causal model based on linear regression of the form

    Yt = A0 + A1 f1t + A2 f2t

    What are the expected signs of A0 , A1 , and A2 , and why? Give the estimated
    values of A0 , A1 , and A2 .
    (b) Is this a good model? (Answer by looking at the R2 value of your regression,
    deciding if all the factors you used are significant, and analyzing the residual
    (error) plots against the factors.) If you don’t think this is a good model can you
    think of adding or deleting factors to make it better?
    (c) We expect the following shipments to arrive in the next 3 days.

    Number of drums 6 12 25
    Total weight 12 25 52

    Forecast how much time it will take to handle the shipments on each day.

    You might also like