Demand Forecasting in a

Supply Chain

Chopra and Meindl, Chapter 4

Learning Objectives
 Phases of supply chain decisions
 Identify components of a demand forecast
 Time series forecasting
 Estimate forecast error
 Aggregate planning in the supply chain
Outline
 The role of forecasting in a supply chain
 Characteristics of forecasts
 Components of forecasts and forecasting methods
 Basic approach to demand forecasting
 Time series forecasting methods
 Measures of forecast error
 Forecasting demand at NaturalGas.com
Phases of Supply Chain Decisions
 Strategy or design: Forecast
 Planning: Forecast
 Operation Actual demand
Role of Forecasting
in a Supply Chain
 The basis for all strategic and planning decisions in a
supply chain
 Used for both push and pull processes
 Examples:
– Production: scheduling, inventory, aggregate planning
– Marketing: sales force allocation, promotions, new production
introduction
– Finance: Plant/equipment investment, budgetary planning
– Personnel: Workforce planning, hiring, layoffs
 All of these decisions are interrelated
Characteristics of Forecasts
 Forecasts are always wrong. Should include
expected value and measure of error.
 Long-term forecasts are less accurate than short-
term forecasts (forecast horizon is important)
 Aggregate forecasts are more accurate than
disaggregate forecasts
Forecasting Methods
 Qualitative: primarily subjective; rely on judgment
and opinion
 Time Series: use historical demand only
– Static
– Adaptive
 Causal: use the relationship between demand and
some other factor to develop forecast
 Simulation
– imitate consumer choices that give rise to demand
– can combine time series and causal methods
 Components of an observation
Observed demand (O) =
Systematic component (S) + Random component (R)
Level (current deseasonalized demand)

Trend (growth or decline in demand)

Seasonality (predictable seasonal fluctuation)

•Systematic component: Expected value of demand
•Random component: The part of the forecast that deviates
from the systematic component
•Forecast error: difference between forecast and actual demand
Time Series Forecasting
Quarter Demand Dt
II, 1998 8000
III, 1998 13000 Forecast demand for the
IV, 1998 23000 next four quarters.
I, 1999 34000
II, 1999 10000
III, 1999 18000
IV, 1999 23000
I, 2000 38000
II, 2000 12000
III, 2000 13000
IV, 2000 32000
I, 2001 41000
Time Series Forecasting

50,000
40,000
30,000
20,000
10,000
0

7, 2 7, 3 7, 4 8, 1 8, 2 8, 3 8, 4 9, 1 9, 2 9, 3 9, 4 0, 1
9 9 9 9 9 9 9 9 9 9 9 0
Forecasting methods
 Static
 Adaptive
– Moving average
– Simple exponential smoothing
– Holt’s model (with trend)
– Winter’s model (with trend and seasonality)
Basic Approach to Demand Forecasting
 Understand the objectives of forecasting
 Integrate demand planning and forecasting
 Identify major factors that influence the demand forecast
 Understand and identify customer segments
 Determine the appropriate forecasting technique
 Establish performance and error measures for the forecast
Time Series
Forecasting Methods
 Goal is to predict systematic component of
demand
– Multiplicative: (level)(trend)(seasonal factor)
– Additive: level + trend + seasonal factor
– Mixed: (level + trend)(seasonal factor)
 Static methods
 Adaptive forecasting
Static Methods
 Assume a mixed model
Systematic component = (level + trend)(seasonal factor)
Ft+l = [L + (t + l)T]St+l
= forecast in period t for demand in period t + l
L = estimate of level for period 0
T = estimate of trend
St = estimate of seasonal factor for period t
Dt = actual demand in period t
Ft = forecast of demand in period t
Static Methods
 Estimating level and trend
 Estimating seasonal factors
Estimating Level and Trend
 Before estimating level and trend, demand data
must be deasonalized
 Deseasonalized demand = demand that would
have been observed in the absence of seasonal
fluctuations
 Periodicity (p)
– the number of periods after which the seasonal cycle
repeats itself
– for demand at NaturalGas(Table 4.1, Figure 4.1) p = 4
Time Series Forecasting (Table 4.1)
Quarter Demand Dt
II, 1998 8000
III, 1998 13000 Forecast demand for the
IV, 1998 23000 next four quarters.
I, 1999 34000
II, 1999 10000
III, 1999 18000
IV, 1999 23000
I, 2000 38000
II, 2000 12000
III, 2000 13000
IV, 2000 32000
I, 2001 41000
Time Series Forecasting (Figure 4.1)

50,000
40,000
30,000
20,000
10,000
0

7, 2 7, 3 7, 4 8, 1 8, 2 8, 3 8, 4 9, 1 9, 2 9, 3 9, 4 0, 1
9 9 9 9 9 9 9 9 9 9 9 0
Estimating Level and Trend
 Before estimating level and trend, demand data
must be deasonalized
 Deseasonalized demand = demand that would
have been observed in the absence of seasonal
fluctuations
 Periodicity (p)
– the number of periods after which the seasonal cycle
repeats itself
– for demand at NaturalGas(Table 4.1, Figure 4.1) p = 4
Deseasonalizing demand

[Dt-(p/2) + Dt+(p/2) +  2Di] / 2p for p even

Dt = (sum is from i = t+1-(p/2) to t+1+(p/2))

 Di / p for p odd
(sum is from i = t-(p/2) to t+(p/2)), p/2 truncated to lower integer
Deasonalizing demand
For the example, p = 4 is even
For t = 3:
D3 = {D1 + D5 + Sum(i=2 to 4) [2Di]}/8
= {8000+10000+[(2)(13000)+(2)(23000)+(2)(34000)]}/8
= 19750
D4 = {D2 + D6 + Sum(i=3 to 5) [2Di]}/8
= {13000+18000+[(2)(23000)+(2)(34000)+(2)(10000)]/8
= 20625
Deasonalizing demand
Then include trend
Dt = L + tT
where Dt = deseasonalized demand in period t
L = level (deasonalized demand at period 0)
T = trend (rate of growth of deasonalized demand)
Trend is determined by linear regression using deseasonalized
demand as the dependent variable and period as the independent
variable (can be done in Excel)
In the example, L = 18,439 and T = 524
Time Series of Demand (Figure 4.3)

50000
40000
Demand

30000 Dt
20000 Dt-bar

10000
0
1 2 3 4 5 6 7 8 9 10 11 12
Period
Estimating Seasonal Factors
Use the previous equation to calculate deasonalized
demand for each period
St = Dt / Dt = seasonal factor for period t
In the example,
D2 = 18439 + (524)(2) = 19487 D2 = 13000
S2 = 13000/19487 = 0.67
The seasonal factors for the other periods are calculated
in the same manner
Estimating Seasonal Factors (Fig. 4.4)
t Dt Dt-bar S-bar
1 8000 18963 0.42 = 8000/18963
2 13000 19487 0.67 = 13000/19487
3 23000 20011 1.15 = 23000/20011
4 34000 20535 1.66 = 34000/20535
5 10000 21059 0.47 = 10000/21059
6 18000 21583 0.83 = 18000/21583
7 23000 22107 1.04 = 23000/22107
8 38000 22631 1.68 = 38000/22631
9 12000 23155 0.52 = 12000/23155
10 13000 23679 0.55 = 13000/23679
11 32000 24203 1.32 = 32000/24203
12 41000 24727 1.66 = 41000/24727
Estimating seasonal factors
The overall seasonal factor for a “season” is then obtained by
averaging all of the factors for a “season”
If there are r seasonal cycles, for all periods of the form pt+i,
1<i<p, the seasonal factor for season i is
Si = [Sum(j=0 to r-1) Sjp+i]/ r
In the example, there are 3 seasonal cycles in the data and p=4, so
S1 = (0.42+0.47+0.52)/3 = 0.47
S2 = (0.67+0.83+0.55)/3 = 0.68
S3 = (1.15+1.04+1.32)/3 = 1.17
S4 = (1.66+1.68+1.66)/3 = 1.67
Estimating the forecast
Using the original equation, we can forecast the next
four periods of demand:

F13 = (L+13T)S1 = [18439+(13)(524)](0.47) = 11868

F14 = (L+14T)S2 = [18439+(14)(524)](0.68) = 17527
F15 = (L+15T)S3 = [18439+(15)(524)](1.17) = 30770
F16 = (L+16T)S4 = [18439+(16)(524)](1.67) = 44794
Adaptive Forecasting
 The estimates of level, trend, and seasonality are
adjusted after each demand observation
 General steps in adaptive forecasting
 Moving average
 Simple exponential smoothing
 Trend-corrected exponential smoothing (Holt’s
model)
 Trend- and seasonality-corrected exponential
smoothing (Winter’s model)
Basic formula for adaptive forecasting
Ft+1 = (Lt + lT)St+1 = forecast for period t+l in period t
Lt = Estimate of level at the end of period t
Tt = Estimate of trend at the end of period t
St = Estimate of seasonal factor for period t
Ft = Forecast of demand for period t (made period t-1 or earlier)
Dt = Actual demand observed in period t
Et = Forecast error in period t
At = Absolute deviation for period t = |Et|
MAD = Mean Absolute Deviation = average value of At
General Steps in Adaptive Forecasting
 Initialize: Compute initial estimates of level (L0), trend (T0), and
seasonal factors (S1,…,Sp). This is done as in static forecasting
 Forecast: forecast demand for period t+1 using the general
equation
 Estimate error: Compute error Et+1 = Ft+1- Dt+1
 Modify estimates: Modify the estimates of level (Lt+1), trend
(Tt+1), and seasonal factor (St+p+1), given the error Et+1 in the
forecast
 Repeat steps 2, 3, and 4 for each subsequent period
Moving Average
 Used when demand has no observable trend or seasonality
 Systematic component of demand = level
 The level in period t is the average demand over the last N periods (the N-
period moving average)
 Current forecast for all future periods is the same and is based on the
current estimate of the level
Lt = (Dt + Dt-1 + … + Dt-N+1) / N
Ft+1 = Lt and Ft+n = Lt
After observing the demand for period t+1, revise the estimates as follows:
Lt+1 = (Dt+1 + Dt + … + Dt-N+2) / N
Ft+2 = Lt+1
 Moving Average Example
From NaturalGas example (Table 4.1)
At the end of period 4, what is the forecast demand for periods 5 through
8 using a 4-period moving average?
L4 = (D4+D3+D2+D1)/4 = (34000+23000+13000+8000)/4 = 19500
F5 = 19500 = F6 = F7 = F8
Observe demand in period 5 to be D5 = 10000
Forecast error in period 5, E5 = F5 - D5 = 19500 - 10000 = 9500
Revise estimate of level in period 5:
L5 = (D5+D4+D3+D2)/4 = (10000+34000+23000+13000)/4 = 20000
F6 = L5 = 20000
Simple Exponential Smoothing
 Used when demand has no observable trend or seasonality
 Systematic component of demand = level
 Initial estimate of level, L0, assumed to be the average of all historical data
L0 = [Sum(i=1 to n)Di]/n
Current forecast for all future periods is equal to the current estimate of the
level and is given as follows:
Ft+1 = Lt and Ft+n = Lt
After observing demand Dt+1, revise the estimate of the level:
Lt+1 = Dt+1 + (1-)Lt
Lt+1 = Sum(n=0 to t+1)[(1-)nDt+1-n ]
 Simple Exponential Smoothing Example
From NaturalGas data, forecast demand for period 1 using exponential smoothing
L0 = average of all 12 periods of data
= Sum(i=1 to 12)[Di]/12 = 22083
F1 = L0 = 22083
Observed demand for period 1 = D1 = 8000
Forecast error for period 1, E1, is as follows:
E1 = F1 - D1 = 22083 - 8000 = 14083
Assuming  = 0.1, revised estimate of level for period 1:
L1 = D1 + (1-)L0 = (0.1)(8000) + (0.9)(22083) = 20675
F2 = L1 = 20675
Note that the estimate of level for period 1 is lower than in period 0
Trend-Corrected Exponential Smoothing
(Holt’s model)
 Appropriate when the demand is assumed to have a level and trend
in the systematic component of demand but no seasonality
 Obtain initial estimate of level and trend by running a linear
regression of the following form:
Dt = at + b
T0 = a
L0 = b
In period t, the forecast for future periods is expressed as follows:
Ft+1 = Lt + Tt
Ft+n = Lt + nTt
Trend-Corrected Exponential Smoothing
(Holt’s model)
After observing demand for period t, revise the estimates for level and trend as
follows:
Lt+1 = Dt+1 + (1-)(Lt + Tt)
Tt+1 = (Lt+1 - Lt) + (1-)Tt
 = smoothing constant for level
 = smoothing constant for trend
Example: NaturalGas demand data. Forecast demand for period 1 using
Holt’s model (trend corrected exponential smoothing)
Using linear regression,
L0 = 12015 (linear intercept)
T0 = 1549 (linear slope)
Holt’s model example continued
Forecast for period 1:
F1 = L0 + T0 = 12015 + 1549 = 13564
Observed demand for period 1 = D1 = 8000
E1 = F1 - D1 = 13564 - 8000 = 5564
Assume  = 0.1,  = 0.2
L1 = D1 + (1-)(L0+T0) = (0.1)(8000) + (0.9)(13564) = 13008
T1 = (L1 - L0) + (1-)T0 = (0.2)(13008 - 12015) + (0.8)(1549)
= 1438
F2 = L1 + T1 = 13008 + 1438 = 14446
F5 = L1 + 4T1 = 13008 + (4)(1438) = 18760
Trend- and Seasonality-Corrected
Exponential Smoothing
 Appropriate when the systematic component of demand is
assumed to have a level, trend, and seasonal factor
 Systematic component = (level+trend)(seasonal factor)
 Assume periodicity p
 Obtain initial estimates of level (L0), trend (T0), seasonal factors
(S1,…,Sp) using procedure for static forecasting
 In period t, the forecast for future periods is given by:
Ft+1 = (Lt+Tt)(St+1) and Ft+n = (Lt + nTt)St+n
Trend- and Seasonality-Corrected
Exponential Smoothing continued
After observing demand for period t+1, revise estimates for level, trend, and
seasonal factors as follows:
Lt+1 = a(Dt+1/St+1) + (1-a)(Lt+Tt)
Tt+1 = b(Lt+1 - Lt) + (1-b)Tt
St+p+1 = (Dt+1/Lt+1) + (1-)St+1
 = smoothing constant for level
 = smoothing constant for trend
 = smoothing constant for seasonal factor
Example: NaturalGas data. Forecast demand for period 1 using Winter’s model
Initial estimates of level, trend, and seasonal factors are obtained as in the static
forecasting case
Trend- and Seasonality-Corrected
Exponential Smoothing example continued
L0 = 18439 T0 = 524 S1=0.47, S2=0.68, S3=1.17, S4=1.67
F1 = (L0 + T0)S1 = (18439+524)(0.47) = 8913
The observed demand for period 1 = D1 = 8000
Forecast error for period 1 = E1 = F1-D1 = 8913 - 8000 = 913
Assume  = 0.1, =0.2, =0.1; revise estimates for level and trend for
period 1 and for seasonal factor for period 5
L1 = (D1/S1)+(1-)(L0+T0) = (0.1)(8000/0.47)+(0.9)(18439+524)=18769
T1 = (L1-L0)+(1-)T0 = (0.2)(18769-18439)+(0.8)(524) = 485
S5 = (D1/L1)+(1-)S1 = (0.1)(8000/18769)+(0.9)(0.47) = 0.47

F2 = (L1+T1)S2 = (18769 + 485)(0.68) = 13093

Measures of Forecast Error
 Forecast error = Et = Ft - Dt
Mean squared error (MSE)
MSEn = (Sum(t=1 to n)[Et2])/n
 Absolute deviation = At = |Et|
Mean absolute deviation (MAD)
MADn = (Sum(t=1 to n)[At])/n
 = 1.25MAD
Measures of Forecast Error
Mean absolute percentage error (MAPE)
MAPEn = (Sum(t=1 to n)[|Et/ Dt|100])/n
 Bias
– Shows whether the forecast consistently under- or overestimates demand;
should fluctuate around 0
biasn = Sum(t=1 to n)[Et]
 Tracking signal
– should be within the range of +6
– otherwise, possibly use a new forecasting method
TSt = bias / MADt
Forecasting Demand at NaturalGas.com
 Moving average
 Simple exponential smoothing
 Trend-corrected exponential smoothing
 Trend- and seasonality-corrected exponential
smoothing