+ All Categories
Home > Documents > the Foecasting

the Foecasting

Date post: 02-Jun-2018
Category:
Upload: rashadafaneh
View: 214 times
Download: 0 times
Share this document with a friend

of 36

Transcript
  • 8/11/2019 the Foecasting

    1/36

    Chapter 8 - Forecasting

    Operations Managementby

    R. Dan Reid & Nada R. Sanders2nd Edition Wiley 2005

  • 8/11/2019 the Foecasting

    2/36

    Decisions that Need ForecastsWhich markets to pursue?What products to produce?How many people to hire?How many units to purchase?

    How many units to produce? And so on

  • 8/11/2019 the Foecasting

    3/36

    Common Characteristics of

    ForecastingForecasts are rarely perfect

    Forecasts are more accurate foraggregated data than for individualitems

    Forecast are more accurate for shorterthan longer time periods

  • 8/11/2019 the Foecasting

    4/36

    Forecasting StepsWhat needs to be forecast?

    Level of detail, units of analysis & time horizonrequired

    What data is available to evaluate?Identify needed data & whether its available

    Select and test the forecasting model

    Cost, ease of use & accuracyGenerate the forecastMonitor forecast accuracy over time

  • 8/11/2019 the Foecasting

    5/36

    Types of Forecasting ModelsQualitative (technological) methods:

    Forecasts generated subjectively by theforecaster

    Quantitative (statistical) methods:Forecasts generated through mathematicalmodeling

  • 8/11/2019 the Foecasting

    6/36

    Qualitative MethodsType Characteristics Strengths Weaknesses

    Executiveopinion

    A group of managersmeet & come up with

    a forecast

    Good for strategic ornew-product

    forecasting

    One person's opinioncan dominate the

    forecast

    Marketresearch

    Uses surveys &interviews to identifycustomer preferences

    Good determinant ofcustomer preferences

    It can be difficult todevelop a goodquestionnaire

    Delphimethod

    Seeks to develop aconsensus among agroup of experts

    Excellent forforecasting long-termproduct demand,technological

    Time consuming todevelop

  • 8/11/2019 the Foecasting

    7/36

    Statistical ForecastingTime Series Models:

    Assumes the future will follow same patterns as

    the past

    Causal Models:Explores cause-and-effect relationshipsUses leading indicators to predict the futureE.g. housing starts and appliance sales

  • 8/11/2019 the Foecasting

    8/36

    Composition

    of Time Series DataData = historic pattern + randomvariationHistoric pattern may include:

    Level (long-term average)Trend

    SeasonalityCycle

  • 8/11/2019 the Foecasting

    9/36

    Time Series Patterns

  • 8/11/2019 the Foecasting

    10/36

    Methods of Forecasting the LevelNave ForecastingSimple MeanMoving AverageWeighted Moving Average

    Exponential Smoothing

  • 8/11/2019 the Foecasting

    11/36

    Time Series ProblemDetermine forecast forperiods 11

    Nave forecast Simple average

    3- and 5-period movingaverage

    3-period weighted movingaverage with weights 0.5,0.3, and 0.2

    Exponential smoothing

    with alpha=0.2 and 0.5

    Period Orders1 1222 913 1004 775 1156 58

    7 758 1289 111

    10 8811

  • 8/11/2019 the Foecasting

    12/36

    Time Chart of Orders Data

    0

    20

    40

    60

    80

    100

    120

    140

    1 2 3 4 5 6 7 8 9 10

  • 8/11/2019 the Foecasting

    13/36

  • 8/11/2019 the Foecasting

    14/36

    Simple Average (Mean)Next periods forecast = average of allhistorical data

    n A A A

    F t t t t .............21

    1

  • 8/11/2019 the Foecasting

    15/36

  • 8/11/2019 the Foecasting

    16/36

    The Effect of the Parameter N

    A smaller N makes the forecast more

    responsive A larger N makes the forecast morestable

  • 8/11/2019 the Foecasting

    17/36

    Weighted Moving Average

    1.........

    .........

    21

    11211

    N

    N t N t t t

    C C C

    where

    AC AC AC F

  • 8/11/2019 the Foecasting

    18/36

  • 8/11/2019 the Foecasting

    19/36

    The Effect of the Parameter

    A smaller makes the forecast more

    stable A larger makes the forecast moreresponsive

  • 8/11/2019 the Foecasting

    20/36

    Time Series Problem Solution

    Simple Simple Weighted Exponential ExponentialNave Simple Moving Moving Moving Smoothing Smoothing

    Period Orders (A) Forecast Average Average (N=3) Average(N=5) Average (N=3) ( = 0.2) ( = 0.5)

    1 122 122 1222 91 122 122 122 1223 100 91 107 116 1074 77 100 104 104 102 113 1045 115 77 98 89 87 106 916 58 115 101 97 101 101 108 1037 75 58 94 83 88 79 98 81

    8 128 75 91 83 85 78 93 789 111 128 96 87 91 98 100 10310 88 111 97 105 97 109 102 107

    11 88 97 109 92 103 99 98

  • 8/11/2019 the Foecasting

    21/36

  • 8/11/2019 the Foecasting

    22/36

    Tracking Forecast ErrorOver Time

    Mean Absolute Deviation (MAD): A good measure of the actual errorin a forecast

    Mean Square Error (MSE):Penalizes extreme errors

    Tracking SignalExposes bias (positive or negative)

    n

    forecast actual MAD

    2

    actual - forecastMSE

    n

    MAD

    TS forecast-actual

  • 8/11/2019 the Foecasting

    23/36

  • 8/11/2019 the Foecasting

    24/36

    Forecasting Trends

    Trend-adjusted exponential smoothingThree step process:

    Smooth the level of the series:

    Smooth the trend:

    Calculate the forecast including trend:

    ))(1( 11 t t t t T S AS

    11 )1()( t t t t T S S T

    t t t T S FIT 1

  • 8/11/2019 the Foecasting

    25/36

    Forecasting trend problem: a company uses exponential smoothing with trend toforecast usage of its lawn care products. At the end of July the company wishes toforecast sales for August. July demand was 62. The trend through June has been 15additional gallons of product sold per month. Average sales have been 57 gallonsper month. The company uses alpha+0.2 and beta +0.10. Forecast for August.

    Smooth the level of the series:

    Smooth the trend:

    Forecast including trend:

    14.8150.957700.1)T(1)S(ST 1t1ttJuly

    7015570.8620.2)T)(S(1AS 1t1ttJuly

    gallons84.814.870TSFIT ttAugust

  • 8/11/2019 the Foecasting

    26/36

    Adjusting for Seasonality

    Calculate the average demand per season E.g.: average quarterly demand

    Calculate a seasonal index for each season ofeach year:

    Divide the actual demand of each season by theaverage demand per season for that year

    Average the indexes by season E.g .: take the average of all Spring indexes, thenof all Summer indexes, ...

  • 8/11/2019 the Foecasting

    27/36

    Adjusting for Seasonality

    Forecast demand for the next year & divideby the number of seasons

    Use regular forecasting method & divide by fourfor average quarterly demand

    Multiply next years average seasonal demandby each average seasonal index

    Result is a forecast of demand for each season ofnext year

  • 8/11/2019 the Foecasting

    28/36

  • 8/11/2019 the Foecasting

    29/36

    Seasonality Problem: Solution

    Quarter Year 1 SeasonalIndex

    Year 2 SeasonalIndex

    Avg.Index

    Year3

    Fall 24000 1.20 26000 1.24 1.22 27450Winter 23000 1.15 22000 1.05 1.10 24750Spring 19000 0.95 19000 0.90 0.93 20925

    Summer 14000 0.70 17000 0.81 0.76 17100

    Total 80000 4.00 84000 4.00 4.01 90000 Average 20000 21000 22500

  • 8/11/2019 the Foecasting

    30/36

  • 8/11/2019 the Foecasting

    31/36

    Linear Regression

    X X X

    Y X XY b

    2

    Identify dependent (y) andindependent (x) variablesSolve for the slope of the

    line

    Solve for the y intercept

    Develop your equation forthe trend line

    Y=a + bX

    XbYa

    22 XnX

    YXnXYb

  • 8/11/2019 the Foecasting

    32/36

  • 8/11/2019 the Foecasting

    33/36

    How Good is the Fit?Correlation coefficient (r) measures the direction and strength of the linearrelationship between two variables. The closer the r value is to 1.0 the betterthe regression line fits the data points.

    Coefficient of determination ( ) measures the amount of variation in thedependent variable about its mean that is explained by the regression line.

    Values of ( ) close to 1.0 are desirable.

    .788.982r

    .88858987,1654*(205)-4(10,533)

    589205)(30,282(4)r

    YYn*XXnYXXYnr

    22

    22

    22

    22

    2r

    2r

  • 8/11/2019 the Foecasting

    34/36

    Factors for Selecting aForecasting Model

    The amount & type of available dataDegree of accuracy requiredLength of forecast horizonPresence of data patterns

  • 8/11/2019 the Foecasting

    35/36

    Forecasting Software

    SpreadsheetsMicrosoft Excel, Quattro Pro, Lotus 1-2-3

    Limited statistical analysis of forecast dataStatistical packages

    SPSS, SAS, NCSS, MinitabForecasting plus statistical and graphics

    Specialty forecasting packagesForecast Master, Forecast Pro, Autobox, SCA

  • 8/11/2019 the Foecasting

    36/36


Recommended