Forecasting

Lesson 4.1

Forecasting

Managers are always trying to reduce uncertainty and make better estimates of what will happen in the future This is the main purpose of forecasting Some firms use subjective methods Seat-of-the pants methods, intuition, experience There are also several quantitative techniques Moving averages, exponential smoothing, trend projections, least squares regression analysis

16 October 2010

Operations Research - Asia Pacific College

Steps to Forecasting

Eight steps to forecasting: 1. Determine the use of the forecastwhat objective are we trying to obtain? 2. Select the items or quantities that are to be forecasted 3. Determine the time horizon of the forecast 4. Select the forecasting model or models 5. Gather the data needed to make the forecast 6. Validate the forecasting model 7. Make the forecast 8. Implement the results
Operations Research - Asia Pacific College
3

16 October 2010

Steps to Forecasting

These steps are a systematic way of initiating, designing, and implementing a forecasting system When used regularly over time, data is collected routinely and calculations performed automatically There is seldom one superior forecasting system Different organizations may use different techniques Whatever tool works best for a firm is the one they should use

16 October 2010

Operations Research - Asia Pacific College

Forecasting Models
Forecasting Techniques Qualitative Models Delphi Methods Jury of Executive Opinion Sales Force Composite Consumer Market Survey Time-Series Methods Moving Average Exponential Smoothing Trend Projections Causal Methods Regression Analysis Multiple Regression

Decomposition

16 October 2010

Operations Research - Asia Pacific College

Time-Series Models

Time-series models attempt to predict the future based on the past Common time-series models are

Moving average Exponential smoothing Trend projections Decomposition

Regression analysis is used in trend projections and one type of decomposition model
Operations Research - Asia Pacific College
6

16 October 2010

Causal Models

Causal models use variables or factors that might influence the quantity being forecasted The objective is to build a model with the best statistical relationship between the variable being forecast and the independent variables Regression analysis is the most common technique used in causal modeling

16 October 2010

Operations Research - Asia Pacific College

Qualitative Models

Qualitative models incorporate judgmental or subjective factors Useful when subjective factors are thought to be important or when accurate quantitative data is difficult to obtain Common qualitative techniques are

Delphi method Jury of executive opinion Sales force composite Consumer market surveys
Operations Research - Asia Pacific College
8

16 October 2010

Qualitative Models

Delphi Method an iterative group process where (possibly geographically dispersed) respondents provide input to decision makers Jury of Executive Opinion collects opinions of a small group of high-level managers, possibly using statistical models for analysis Sales Force Composite individual salespersons estimate the sales in their region and the data is compiled at a district or national level Consumer Market Survey input is solicited from customers or potential customers regarding their purchasing plans
Operations Research - Asia Pacific College
9

16 October 2010

Time-Series Forecasting Models

A time series is a sequence of evenly spaced events Time-series forecasts predict the future based solely of the past values of the variable Other variables are ignored

16 October 2010

Operations Research - Asia Pacific College

10

Decomposition of a Time-Series

A time series typically has four components 1. Trend (T) is the gradual upward or downward movement of the data over time 2. Seasonality (S) is a pattern of demand fluctuations above or below trend line that repeats at regular intervals 3. Cycles (C) are patterns in annual data that occur every several years 4. Random variations (R) are blips in the data caused by chance and unusual situations
Operations Research - Asia Pacific College
11

16 October 2010

Decomposition of a Time-Series
Demand for Product or Service
Trend Component Seasonal Peaks Actual Demand Line Average Demand over 4 Years

Year 1

Year 2

Time

Year 3

Year 4

16 October 2010

Operations Research - Asia Pacific College

12

Measures of Forecast Accuracy

We compare forecasted values with actual values to see how well one model works or to compare models

Forecast error = Actual value Forecast value

One measure of accuracy is the mean absolute deviation (MAD)

forecast error MAD n

This means the better model is the one with the lower MAD value.

16 October 2010

Operations Research - Asia Pacific College

13

Moving Averages
Moving averages can be used when data is

relatively steady over time The next forecast is the average of the most recent n data values from the time series This methods tends to smooth out shortterm irregularities in the data series
Moving average forecast Sum of demands in previous n periods n

16 October 2010

Operations Research - Asia Pacific College

14

Moving Averages
Mathematically

Ft 1

Yt Yt 1 ... Yt n1 n

where

Ft 1 = forecast for time period t + 1 Yt = actual value in time period t n = number of periods to average

16 October 2010

Operations Research - Asia Pacific College

15

Example 4.1 Del Monte Catsup I

Del Monte Catsup wants to forecast the sales of catsup on a monthly basis They have collected data for the past year They are using a three-month moving average to forecast sales (n = 3)

16 October 2010

Operations Research - Asia Pacific College

16

Example 4.1 Del Monte Catsup I

MONTH January February March April May June ACTUAL SALES ($000) 8 15 17 16 16 23 THREE-MONTH MOVING AVERAGE

(08 + 15 + 17)/3 = 13.33 (15 + 17 + 16)/3 = 16.00 (17 + 16 + 16)/3 = 16.33

July
August September October November December January
16 October 2010

20
25 28 21 28 29

(16 + 16 + 23)/3 = 18.33

(16 + 23 + 20)/3 = 19.67 (23 + 20 + 25)/3 = 22.67 (20 + 25 + 28)/3 = 24.33 (25 + 28 + 21)/3 = 24.67 (28 + 21 + 28)/3 = 25.67 (21 + 28 + 29)/3 = 26.00
Operations Research - Asia Pacific College
17

Weighted Moving Averages

Weighted moving averages use weights to put

more emphasis on recent periods Often used when a trend or other pattern is emerging
Ft 1 ( Weight in period i )( Actual value in period) ( Weights)

Mathematically

Ft 1
where

w1Yt w2Yt 1 ... wnYt n1 w1 w2 ... wn

wi = weight for the ith observation

16 October 2010

Operations Research - Asia Pacific College

18

Example 4.2 Del Monte Catsup II

Del Monte Catsup decides to try a weighted

moving average model to forecast sales They decide on the following weighting scheme
WEIGHTS APPLIED 3 2 1 6 PERIOD Last month Two months ago Three months ago

3 x Sales last month + 2 x Sales two months ago + 1 X Sales three months ago

Sum of the weights

16 October 2010

Operations Research - Asia Pacific College

19

Example 4.2 Del Monte Catsup II

MONTH January ACTUAL SALES ($000) 8 THREE-MONTH WEIGHTED MOVING AVERAGE

February
March April May June July August September October

15
17 16 16 23 20 25 28 21

[(3 X 17) + (2 X 15) + (08)]/6 = 14.83 [(3 X 16) + (2 X 17) + (15)]/6 = 16.17 [(3 X 16) + (2 X 16) + (17)]/6 = 16.17

[(3 X 23) + (2 X 16) + (16)]/6 = 19.50

[(3 X 20) + (2 X 23) + (16)]/6 = 20.33 [(3 X 25) + (2 X 20) + (23)]/6 = 23.00 [(3 X 28) + (2 X 25) + (20)]/6 = 25.67 [(3 X 21) + (2 X 28) + (25)]/6 = 24.00

November
December January
16 October 2010

28
29

[(3 X 28) + (2 X 21) + (28)]/6 = 25.67

[(3 X 29) + (2 X 28) + (21)]/6 = 27.33
Operations Research - Asia Pacific College
20

Exponential Smoothing

Exponential smoothing is easy to use and requires little record keeping of data It is a type of moving average New forecast = Last periods forecast + (Last periods actual demand Last periods forecast)

Where is a weight (or smoothing constant) with a value between 0 and 1 inclusive

16 October 2010

Operations Research - Asia Pacific College

21

Exponential Smoothing

Mathematically

Ft 1 Ft (Yt Ft )
where
Ft+1 = new forecast (for time period t + 1) Ft = pervious forecast (for time period t) = smoothing constant (0 1) Yt = pervious periods actual demand

The idea is simple the new estimate is the old estimate plus some fraction of the error in the last period
Operations Research - Asia Pacific College
22

16 October 2010

Selecting the Smoothing Constant

Selecting the appropriate value for is key to obtaining a good forecast The objective is always to generate an accurate forecast The general approach is to develop trial forecasts with different values of and select the that results in lowest error.

16 October 2010

Operations Research - Asia Pacific College

23

Example 4.3 Del Monte Catsup III

Del Monte Catsup wants to include past forecasts in the new forecasts. They think the smoothing constant is 0.80

16 October 2010

Operations Research - Asia Pacific College

24

Example 4.3 Del Monte Catsup III

In January, Februarys sales was predicted to be 8 ($000) Actual February sales was 15 ($000) Using a smoothing constant of = 0.80, what is the forecast for March?
New forecast (for March sales) = 8 + 0.8(15 8) = 13.6 ($000)

If actual sales in March was 17 ($000), the April forecast would be

New forecast (for April sales) = 13.6 + 0.8(17 13.6) = 16.3 ($000)

16 October 2010

Operations Research - Asia Pacific College

25

Example 4.3 Del Monte Catsup III

MONTH January February March April ACTUAL SALES ($000) 8 15 17 16 8.0 8.0 + 0.8(15.0 8.0) = 13.6 13.6 + 0.8(17.0 13.6) = 16.3 SALES FORECAST (Using =.80)

May
June July August September October November December January
16 October 2010

16
23 20 25 28 21 28 29

16.3 + 0.8(16.0 16.3) = 16.1

16.1 + 0.8(16.0 16.1) = 16.0 16.0 + 0.8(23.0 16.0) = 21.6 21.6 + 0.8(20.0 21.6) = 20.3 20.3 + 0.8(25.0 20.3) = 24.1 24.1 + 0.8(28.0 24.1) = 27.2 27.2 + 0.8(21.0 27.2) = 22.2 22.2 + 0.8(28.0 22.2) = 26.8 26.8 + 0.8(29.0 26.8) = 28.6
26

Operations Research - Asia Pacific College

Exponential Smoothing with Trend Adjustment

Like all averaging techniques, exponential smoothing does not respond to trends A more complex model can be used that adjusts for trends The basic approach is to develop an exponential smoothing forecast then adjust it for the trend

Forecast including trend (FITt) = New forecast (Ft) + Trend correction (Tt)

16 October 2010

Operations Research - Asia Pacific College

27

Exponential Smoothing with Trend Adjustment

The equation for the trend correction uses a new smoothing constant Tt is computed by
Tt 1 (1 )Tt (Ft 1 Ft )
where Tt+1 = Tt = = Ft+1 = smoothed trend for period t + 1 smoothed trend for preceding period trend smooth constant that we select simple exponential smoothed forecast for period t + 1 Ft = forecast for previous period
Operations Research - Asia Pacific College
28

16 October 2010

Selecting a Smoothing Constant

As with exponential smoothing, a high value of makes the forecast more responsive to changes in trend A low value of gives less weight to the recent trend and tends to smooth out the trend Values are generally selected using a trial-and-error approach Simple exponential smoothing is often referred to as first-order smoothing Trend-adjusted smoothing is called second-order, double smoothing, or Holts method
Operations Research - Asia Pacific College
29

16 October 2010

Example 4.4 Del Monte Catsup IV

Del Monte Catsup in Example 4.3 wants to include a trend adjustment in the new forecasts. They use the smoothing constant =0.80 They think that the trend constant is 0.40 The figures are summarized in the following table. What is the forecast on the monthly sales?

16 October 2010

Operations Research - Asia Pacific College

30

Example 4.4 Del Monte Catsup IV

In January, Februarys sales was predicted to be 8($000) and trend correction is 0. Forecast with trend is then 8 + 0 = 8 ($000). Actual February sales was 15 ($000) Using a smoothing constant of = 0.80, the forecast for March is 13.6 ($000). Using a trend constant of = 0.40, the trend correction is Trend Correction (for March sales) = (1 0.4)(0) + 0.4(13.6 8) = 2.2 ($000)
Forecast with Trend (for March sales) = 13.6 + 2.2 = 15.8 ($000)

16 October 2010

Operations Research - Asia Pacific College

31

Example 4.4 Del Monte Catsup IV

MONTH January February March April May ACTUAL SALES ($000) 8 15 17 16 16 8.0 13.6 16.3 16.1 0.0 0.6(0.0) + 0.4(13.6 - 8.0) = 2.2 0.6(2.2) + 0.4(16.3 -13.6) = 2.4 0.6(2.4) + 0.4(16.1 -16.3) = 1.4 8.0 15.8 18.7 17.5 SALES FORECAST (Using =.80) TREND CORRECTION (Using =.40) FORECAST WITH TREND

June
July August September October November December January

23
20 25 28 21 28 29

16.0
21.6 20.3 24.1 27.2 22.2 26.8 28.6

0.6(1.4) + 0.4(16.0 -16.1) = 0.8

0.6(0.8) + 0.4(21.6 -16.0) = 2.7 0.6(2.7) + 0.4(20.3 -21.6) = 1.1 0.6(1.1) + 0.4(24.1 -20.3) = 2.2 0.6(2.2) + 0.4(27.2 -24.1) = 2.6 0.6(2.6) + 0.4(22.2 -27.2) = -0.5 0.6(-0.5) + 0.4(26.8 -22.2) = 1.6 0.6(1.6) + 0.4(28.6 -26.8) = 1.7

16.8
24.3 21.4 26.3 29.8 21.7 28.4 30.3

16 October 2010

Operations Research - Asia Pacific College

32