Unit II- Forecasting in production and operations management

The Importance of Forecasting

• Governments forecast unemployment, interest rates,

and expected revenues from income taxes for policy
purposes
• Marketing executives forecast demand, sales, and
consumer preferences for strategic planning
• College administrators forecast enrollments to plan
for facilities and for faculty recruitment
• Retail stores forecast demand to control inventory
levels, hire employees and provide training
 Steps in forecasting process
[ Ref : Quantitative Management by Render , Stair , Hanna and
Badri ]

• Determining the use of the forecast –what objective are we

trying to obtain .
• Select the items and quantities that are to be forecasted
• Determine the time horizon of the forecast –is it 1 to 30 days
[ short term ] , 1 month to 1 year [medium term] or more
than one year [ long term] ?
• Select the forecasting models or models
• Gather the data or information needs to make the forecasts
• Validate the forecasting model
• Make the forecast
• Implement the results
Common Approaches to Forecasting

Common
Approaches to
Forecasting
Qualitative Quantitative
forecasting methods forecasting methods
• Used when historical data are
unavailable
• Considered highly subjective Time Causal
and judgmental
• Executive opinion Series
• Consumer survey
Exponential
• Sales force estimate  Naïve approach smoothing
• Delphi technique
 Qualitative approaches

• Unlike Quantitative methods, Qualitative forecasting uses data that can’t be

measured. It relies on opinions and expert advice and is useful for new
companies that don’t have any or much historical data.
• Qualitative methods are useful, but it’s important to take the information into
account in a non-judgmental and unbiased manner.
• The qualitative forecasting method can be separated into the following 4
approaches: executive opinion, Delphi method, sales force estimates, and
consumer surveys.
1) Executive Opinion
• This happens when a group of executives makes decisions regarding the future of
a company.
• Example: Coca-Cola executives use their industry-experience to make an
important decision regarding the marketing campaign for the new year.
•
 Qualitative approaches
2) Delphi Method
• The Delphi method is used as a group of industry professionals give an opinion
regarding the future of the company.
• This opinion is given to another group of experts, who interprets and discusses the
opinion before presenting the ‘final’ opinion to company decision makers. Example: A
manager has all employees anonymously submit their predictions for sales over the
next quarter. He then compiles them, looks them over, and submits a final forecast to
his boss who will create the actual forecast.
3) Sales Force Estimates
• This happens when individual sales employees use their experience and knowledge to
make their own sales forecasts.
• Example: Sales representatives are expecting to close three major deals this month, so
the forecast is adjusted to represent their predictions.
4) Consumer Surveys
• Consumers give their opinions on products. These opinions are considered when
forecasting.
• Example: You hand out surveys to each customer that makes a purchase. Each survey
has the same questions and they’re submitted anonymously.
 Measuring Errors
• Choose the model that gives the smallest
measuring errors
 Sum of squared errors  Mean Absolute Deviation
(SSE) n (MAD)n
SSE  (Yi  Ŷi )2  Y  Ŷ
i i
i1 MAD  i1
n

 Sensitive to outliers
 Not sensitive to extreme
observations
 Residual Analysis
e e

0 0

T T
Random Cyclical effects not
e errors accounted for
e
0 0

T T
Trend not Seasonal effects not
accounted for accounted for
 Time Series -- Meaning
• Time Series deals with statistical data which are
collected, observed or recorded at successive
intervals of time.

• A time series is a set of observations taken at

specified times, usually at equal intervals.

• A time series may be defined by the values of Y1, Y2,

… of a variable Y at times t1, t2,… Thus, Y = f(t)
 Components of time
Secular Trend
series
The secular trend refers to the general tendency of
data to grow or decline over a long period of time.
Seasonal Variations
Seasonal variations occur in the time series due to the rhythmic forces which
occurs in a regular and a periodic manner with in a period of less than one year.
Seasonal variations occur during a period of one year and have the same pattern
year after year.
Cyclical Variations
The rhythmic movements in a time series with a period of oscillation (repeated
again and again in same manner) more than one year is called a cyclical
variation and the period is called a cycle.
Irregular variations
It is also called Erratic, Accidental or Random Variations.
The three variations trend, seasonal and cyclical variations are called as regular
variations, but almost all the time series including the regular variation contain
another variation called as random variation.
 Time Series : Traditional View
• Trend, business cycle and seasonal cycle, are not influenced
by stochastic disturbances and can thus be represented by
deterministic functions of time.
• Stochastic impact is restricted to the residuals, and they do
not contain any systematic movements.
• Traditional View hypothesize two models for time
series –
– One, Additive – where all the components are
behaving in additive manner:
Yt Tt  S t  Ct  Rt
Yt Tt St Ct Rt
– Another, Multiplicative – where all the components are
behaving in multiplicative manner:
 Time-Series Data

• Numerical data obtained at regular time intervals.

• The time intervals can be annually, quarterly, daily,
hourly, etc.
• Example:

Year: 2000 2001 2002 2003 2004

Sales [‘000] : 75.3 74.2 78.5 79.7
80.2
 Time-Series Plot
A time-series plot is a two-
dimensional plot of time series
data
• the vertical axis measures U.S. Inflation Rate
the variable of interest 16.00
14.00
12.00
• the horizontal axis 10.00
Inflation Rate (%)

8.00
corresponds to the time 6.00
periods 4.00
2.00
0.00
1975
1977
1979
1981
1983
1985
1987
1989
1991
1993
1995
1997
1999
2001
Year
 Trend Component
• Long-run increase or decrease over time (overall upward or downward
movement)
• Data taken over a long period of time.
• Trend can be upward or downward
• Trend can be linear or non-linear

d
U pwa r
trend
Sale
s

Tim
 Seasonal Component
• Short-term regular wave-like patterns
• Observed within 1 year
• Often monthly or quarterly

Sales
Summer
Winter
Summer
Winter Spring Fall

Spring Fall

Time (Quarterly)
 Cyclical Component
• Long-term wave-like patterns
• Regularly occur but may vary in length
• Often measured peak to peak or trough
to trough
1
Sale Cycle
s

Year
 Irregular Component

• Unpredictable, random,
“residual” fluctuations
• Due to random variations of
– Nature
– Accidents or unusual events
• “Noise” in the time series
 Multiplicative Time-Series Model for
Annual Data
• Used primarily for forecasting
• Observed value in time series is the product of
components

Yi Ti Ci Ii

where Ti = Trend value at year i
Ci = Cyclical value at year i
Ii = Irregular (random)
value at year i
 Multiplicative Time-Series Model with
a Seasonal Component
• Used primarily for forecasting
• Allows consideration of seasonal variation

Yi Ti Si Ci Ii

where Ti = Trend value at time i
Si = Seasonal value at time i
Ci = Cyclical value at time i
Ii = Irregular (random) value at time i
 Moving Averages

• Used for smoothing

• A series of arithmetic means over time
• Result dependent upon choice of L (length of
period for computing means)
• Examples:
– For a 5 year moving average, L = 5
– For a 7 year moving average, L = 7
– Etc.
 Moving Averages
(continued)
• Example: Five-year moving average

– First average:
Y1  Y2  Y3  Y4  Y5
MA(5) 
5

– Second average:
Y2  Y3  Y4  Y5  Y6
MA(5) 
5

– etc.
 Example: Annual Data
Sales (‘000
Year units)
1 Annual Sales
2 23
3 40 60
4 25 50
5
6
27
32
40 …
30
Sales
7 48
8 33 20
9 37 10
10 37
11 50
0 …
1 2 3 4 5 6 7 8 9 10 11
40 Year
etc… etc…
Calculating Moving Averages
Avera 5-Year
ge Moving
Year Sales Year Average
1 2  3  4  5
1 23 3 29.4 3
5
2 40 4 34.4
3 25 5 33.0 23  40  25  27  32
29.4 
4 27 5
6 35.4
5 32 7 37.4
6 48 8 41.0
7 33 9 39.4
8 37 etc … …
9 37 …
10 50 • Each moving average is for a
11 40 consecutive block of 5 years
 Annual vs. Moving Average

• The 5-year moving Annual vs. 5-Year Moving Average

average smoothes the 60

data and shows the 50
underlying trend 40
30
Sales

20
10
0
1 2 3 4 5 6 7 8 9 10 11
Year

Annual 5-Year Moving Average

 Exponential Smoothing
• A weighted moving average
– Weights decline exponentially
– Most recent observation weighted most
• Used for smoothing and short term forecasting (often one
period into the future).
• The weight (smoothing coefficient) is W
– Subjectively chosen
– Range from 0 to 1
– Smaller W gives more smoothing, larger W gives less smoothing
• The weight is:
– Close to 0 for smoothing out unwanted cyclical and irregular
components
– Close to 1 for forecasting
 Exponential Smoothing Model

 Exponential smoothing model

E1 Y1
Ei WYi  (1  W )Ei 1
For i = 2, 3, 4, …
where:
Ei = exponentially smoothed value for period i
Ei-1 = exponentially smoothed value already
computed for period i - 1
Yi = observed value in period i
W = weight (smoothing
Chap 16-26
coefficient), 0 < W <
1
 Exponential Smoothing Example
• Suppose we use weight W = 0.2
Forecas
Time Sales t from Exponentially
Period (Y ) prior Smoothed Value for
(i) i
period this period (Ei)
(Ei-1) E1 = Y1
23 since no
prior
(.2)(40)+(.8)(23)=26.4
informat
(.2)(25)+(.8) ion
(26.4)=26.12 exists
(.2)(27)+(.8) Ei 
(26.12)=26.296 WYi  (1  W )Ei 1
(.2)(32)+(.8)
(26.296)=27.437
1 23 -- (.2)(48)+(.8)
2 40 23 (27.437)=31.549
3 25 26.4 (.2)(48)+(.8)
4 27 26.12 (31.549)=31.840
 Sales vs. Smoothed Sales
60
• Fluctuations have been
50
smoothed
40

Sales
30
• NOTE: the smoothed value
in this case is generally a 20
little low, since the trend is 10
upward sloping and the
0
weighting factor is only .2 1 2 3 4 5 6 7 8 9 10
Time Period
Sales Smoothed

Forecasting in Time period (i+1).

Ŷi1
The smoothed value in the current period (i) is used as the forecast value for Ei
next period (i + 1) .
 Exponential Smoothing in Excel

• Use tools / data analysis /

exponential smoothing

– The “damping factor” is (1 - W)

 Trend-Based Forecasting
• Estimate a trend line using regression analysis

Time Use time (X) as the independent


Year Period Sales variable:
(X) (Y)
1999 0 20 Ŷ b0  b1X
2000 1 40
2001 2 30
2002 3 50
2003 4 70
2004 5 65
 Trend-Based Forecasting
(continued)
Time
Year Period Sales • The linear trend forecasting equation is:
(X) (Y)
1999 0 20 Ŷi 21.905  9.5714 Xi
2000 1 40 Sales trend
2001 2 30 80
2002 3 50 70
60
2003 4 70 50
40
sales

2004 5 65 30
20
10
0
0 1 2 3 4 5 6

Year
 Trend-Based Forecasting
(continued)
Time • Forecast for time period 6:
Year Period Sales
(X) (y) Ŷ  21.905  9.5714 (6)
1999 0 20
2000 1 40  79.33
2001 2 30
Sales trend
2002 3 50
2003 4 70 80
70
2004 5 65 60
2005 6 ?? 50
40
sales

30
20
10
0
0 1 2 3 4 5 6

Year
 Causal forecast technique
• Causal models incorporate the variables or factors that might
influence the quantity being forecasted into the forecasting model.
For example, daily sales of a cola drink might depend on the
season, the average temperature, the average humidity, whether it
is a weekend or a weekday, and so on.
• Causal forecasting is a strategy that involves the attempt to predict or
forecast future events in the marketplace, based on the range of
variables that are likely to influence the future movement within that
market.
• The idea behind this type of prediction is to determine what type of
impact those anticipated variables will have on consumer demand, the
type of pricing that the market will be able to support in the future,
and what those changes would mean for the future of the company.
 Causal forecasting
• This type of forecasting is helpful to companies in several ways, including
the development of sales and advertising for the upcoming period.
• There are several elements that go into a causal forecasting model.
Typically, the process will begin with an assessment of the market as it
currently stands.
• This will include the current position of the company within that market.
From there, there is a need to identify both dependent and independent
variables that are likely to exert some influence on the direction that the
market will take over a specified period of time.
• Once there is a reasonable projection of what will happen to the market as
a whole, it is possible to apply those same variables and their cumulative
effect to the business operation itself.
• Depending on the outcome of the projections, the company may find it
advantageous to begin increasing production now in anticipation of an
increased demand for its products at a later date.
 Quantitative forecasting – naïve forecast
• The naïve forecasting methods base a projection for a future period on data recorded for a
past period. For example, a naïve forecast might be equal to a prior period's actuals, or the
average of the actuals for certain prior periods
• While Naïve forecasting can be useful in some situations, there are certain situations where
this method of forecasting can be problematic. To demonstrate the naïve method, a spread
sheet with 3 and a half periods worth of sales history is shown below …
• The naïve method of forecasting dictates that we use the previous period to forecast for the
next period. Create a % difference column. This column will show the % of variance between
the Actual Sales column and the forecast. This will show you how accurate the forecast
actually is. Equation used, =IF(Actual Sales=0,0(Naïve Forecast/Actual Sales)-1) , in the
image below:
 Quantitative forecasting – naïve forecast
• What this equation means is that if the forecast is less than the actual sales
within that time period, then the % will be positive. A positive percentage
means that what we actually sold is greater than what we forecast. A
negative percentage means that we sold less than the forecast indicated we
would sell.
• To calculate a naïve forecast simple take the previous month of sales and
plug it in next to the adjacent period. The equation for this
method, =(Previous months actual sales) , is shown below:

Once you've applied the equation, you'll notice that the equation has projected a
positive percentage within 10%. That is a pretty good forecast relatively speaking,
moving in the right direction.
• Thank you