Beruflich Dokumente
Kultur Dokumente
Chain Requirements
Information consists of
1. Historical data on our time series
2. Insight/knowledge and common sense
2000
1800
1600
1400
1200
Sales ($1000)
1000
800
600
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1800
1600
1400
1200
Sales ($1000)
1000
800
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1800
1600
1400
1200
Sales ($1000)
1000
800
600
400
2000
1800
1600
1400
1200
Sales ($1000)
1000
We’ll guess same as last month
800 plus a little more for a possible trend
600
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1800
1600
1400
1200
Sales ($1000)
1000
This is easy, who needs forecasting
800
600
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1800
1600
1400
1200
Sales ($1000)
1000
Continue with our successful method: guess
800 the same as last month plus a little more
for a possible trend
600
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1800
1600
1400
1200
Sales ($1000)
1000
800
600
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1800
1600
1400
1200
Sales ($1000)
1000
Definitely looks like a trend
800
600
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1800
1600
1400
1200
Sales ($1000)
1000
800
600
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1800
1600
1400
1200
Sales ($1000)
1000
Trend might be a tad steeper
800 than I thought
600
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1800
1600
1400
1200
Sales ($1000)
1000
Opps
800
600
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1800
1600
1400
1200
Sales ($1000)
1000
Momentary deviation,
800 trend will continue
600
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1800
1600
1400
1200
Sales ($1000)
1000
See, I told you this was easy!
800
600
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1800
1600
1400
Trend will continue
1200
Sales ($1000)
1000
800
600
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1800
1600
1400
1200
Sales ($1000)
1000
Opps, another momentary fluctuation:
800
600
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1800
1600
1000
800
600
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1800
1600
1400 Oh oh!
1200
Sales ($1000)
1000
800
600
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1800
1600
1400
1200
Sales ($1000)
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1800
1600
1400
1200
Sales ($1000)
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1800
1600
1400
1200
Sales ($1000)
600
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1800
1600
1400
1200
Sales ($1000)
600
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1800
1600
1400
1200
Sales ($1000)
600
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1800
1600
This stuff is easy
1400
1200
Sales ($1000)
1000
800
600
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1800
1600
We have for sure
1400
leveled off
1200
Sales ($1000)
1000
800
600
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1800
1000
800
600
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1800
1600
1000
800
600
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1800
1600
Remarkable turnaround in sales.
1400
New CEO Smithers given credit
1200
Sales ($1000)
1000
800
600
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1800
1600
1400
1000
800
600
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1800
1600
1400
1000
800
600
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1800
1600
1400
1000
800
600
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1800
1600
1400
1000
800
600
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1800
1600
1400
I’ll hedge my bets
1200
Sales ($1000)
1000
800
600
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1200
Sales ($1000)
1000
800
600
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1800
1600
Trend up!
1400
1200
Sales ($1000)
1000
800
600
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1800
1600
Not bad!
1400
1200
Sales ($1000)
1000
800
600
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1800
Revise trend a tad
1600
1400
1200
Sales ($1000)
1000
800
600
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1400
1200
Sales ($1000)
1000
800
600
400
Smithers
200
Smothers Monthly Sales
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1400
1200
Sales ($1000)
1000
800
600
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1800
No big deal,
1600
trend continues
1400
1200
Sales ($1000)
1000
800
(in an unrelated matter
Smithers cashes out
600
stock options)
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1800
1600
1400
1200
Sales ($1000)
1000
800
600
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1800
1600
1400
1200
Sales ($1000)
1000
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1800
1600
1400
1200
Sales ($1000)
1000
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1800
Smithers called
1600 before board
1400
1200
Sales ($1000)
1000
800
600
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1800
1600
1400
1200
Sales ($1000)
1000
800
600
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1200
Sales ($1000)
1000
800
600
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1800
1600
1400
1200
Sales ($1000)
1000
800
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1800
1600
1400
1200
Sales ($1000)
1000
800
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1800
1600
1400
1200
Sales ($1000)
1000
800
600
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
Monthly Sales and Forecast
2000
1400
1200
Sales ($1000)
1000
800
600
400
Monthly Sales
200
Forecast
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Dr.Burcu Ozcam LOG301
Month
What have we learned?
Our Actual sales appears to be a great leading indicator of our forecast
• It is supposed to work the other way around!!!!
If we add up the (absolute value of) our forecast errors, we get 226.2
2500
Sales($1000)
2000
1500
1000
500
0
1
8
15
22
29
36
43
85
50
57
64
71
78
92
99
106
113
120
127
134
141
M o n th
Dr.Burcu Ozcam LOG301
Value of Data
Given data, we can forecast this series quite
accurately.
This assumes stable behavior
Recommend at least 4 - 5 seasons of data.
Monthly demand thus prefers 4 to 5 years of data
With 2 years of data, we are essentially forecasting on
the basis of two points if there is seasonal behavior
1. Even given that the future behaves like the past, there is a limit to how accurate
forecasts can be (or nothing can be predicted with complete accuracy)
• The key issue is: How close will the forecast be to the actual value?
• It is crucial to attempt to quantify the expected accuracy of a forecast
1. The further into the future you attempt to forecast, the greater will be the forecast error.
• Major decisions are often based on long term forecasts. e.g. building a new plant
• Considering risk is even more important in these cases
20
15 Now
10
5
Demand
-5
Past Data Future forecast
-10
-15
-20
1
5
9
13
21
25
29
33
41
57
65
69
73
17
37
45
49
53
61
P e rio d
Dr.Burcu Ozcam LOG301
Fore casts with 50% P re diction Inte rv als
20
15
10
5
Demand
-5
-10
-15
-20
1
5
9
13
21
25
29
33
41
57
65
69
73
17
37
45
49
53
61
P e rio d
Dr.Burcu Ozcam LOG301
Fore casts with 95% P re diction Inte rv als
20
15
10
5
Demand
-5
-10
-15
-20
1
5
9
13
21
25
29
33
41
57
65
69
73
17
37
45
49
53
61
P e rio d
Dr.Burcu Ozcam LOG301
Time-Series Data
Numerical data obtained at regular time
intervals
The time intervals can be annually, quarterly,
daily, hourly, etc.
Example:
Year: 1999 2000 2001 2002 2003
Sales: 75.3 74.2 78.5 79.7 80.2
10.00
the horizontal axis 8.00
corresponds to the 6.00
4.00
time periods 2.00
0.00
1975
1977
1979
1981
1989
1993
1995
1997
1999
2001
1983
1985
1987
1991
Dr.Burcu Ozcam LOG301
Year
Time-Series Components
Time-Series
Sales rd t re nd
U pwa
Sales Sales
Time Time
Downward linear trend Upward nonlinear trend
Sales
Summer
Winter
Spring Fall
Time (Quarterly)
Dr.Burcu Ozcam LOG301
Cyclical Component
Long-term wave-like patterns
Regularly occur but may vary in length
Often measured peak to peak or trough to trough
1 Cycle
Sales
2000 2 40 80
70
2001 3 30 60
2002 4 50 50
sales
40
2003 5 70 30
20
2004 6 65 10
0
0 1 2 3 4 5 6
Dr.Burcu Ozcam LOG301 Year
Trend-Based Forecasting
Forecast for time period 7:
Time
Year Period (t) Sales ŷ = 12.333 + 9.5714 (7)
(y)
1999 1 20 = 79.33Sales
2000 2 40
80
2001 3 30 70
2002 4 50 60
2003 5 70 50
sales
40
2004 6 65 30
2005 7 ?? 20
10
0
0 1 2 3 4 5 6 7
Dr.Burcu Ozcam LOG301 Year
Comparing Forecast Values
to Actual Data
The forecast error or residual is the difference
between the actual value in time t and the forecast
value in time t:
Error in time t:
e t = y t − Ft
MSE =
∑ (y t − Ft ) 2
MAD =
∑ |y t − Ft |
n n
where:
yt = Actual value at time t
Ft = Predicted value at time t
n = Number of time periods
Dr.Burcu Ozcam LOG301
Moving Averages
Used for smoothing
Series of arithmetic means over time
Result dependent upon choice of L (length of
period for computing means)
To smooth out seasonal variation, L should
be equal to the number of seasons
• For quarterly data, L = 4
• For monthly data, L = 12
• First average:
Q1 + Q2 + Q3 + Q4
Moving average 1 =
4
• Second average:
Q2 + Q3 + Q4 + Q5
Moving average 2 =
4
• etc…
1 23 Quarterly Sales
2 40
60
3 25
4 27
50
…
40
5 32
Sales
6 48 30
7 33 20
8 37 10 …
9 37 0
10 50 1 2 3 4 5 6 7 8 9 10 11
11 40 Quarter
etc… etc…
Ft +1 = Ft + α( y t − Ft )
or:
Ft +1 = αy t + (1 − α )Ft
where:
Ft+1 = forecast value for period t + 1
yt = actual value for period t
Ft = forecast value for period t
α = alpha (smoothing constant)
Dr.Burcu Ozcam LOG301
Exponential Smoothing Example
Suppose we use weight α = .2
Quarter (t) Sales Forecast Forecast for next period (Ft+1 )
(yt) from prior
period
F1 = y1 since
1 23 NA 23 no prior
2 40 23 (.2)(40)+(.8)(23)=26.4 information
3 25 26.4 (.2)(25)+(.8)(26.4)=26.12 exists
4 27 26.12 (.2)(27)+(.8)(26.12)=26.296
5 32 26.296 (.2)(32)+(.8)(26.296)=27.437
6 48 27.437 (.2)(48)+(.8)(27.437)=31.549 Ft +1
7 33 31.549 (.2)(48)+(.8)(31.549)=31.840 = αy t + (1 − α)Ft
8 37 31.840 (.2)(33)+(.8)(31.840)=32.872
9 37 32.872 (.2)(37)+(.8)(32.872)=33.697
10 50 33.697 (.2)(50)+(.8)(33.697)=36.958
etc… etc… etc… etc…
y = b 0 + b1x + ε
Variable
y y = b 0 + b1x + ε
Observed Value
of y for xi
εi Slope = b1
Predicted Value Random Error
of y for xi
for this x value
Intercept = b0
xi x
Dr.Burcu Ozcam LOG301
Least Squares Criterion
∑e 2
= ∑ (y −ŷ) 2
= ∑ (y − (b 0 + b1x))
2
b1 =
∑ ( x − x )( y − y )
∑ (x − x) 2
algebraic equivalent:
and
∑ xy − ∑ x∑ y
b1 = n b0 = y − b1 x
∑ x 2
−
( ∑ x ) 2
n
Dr.Burcu Ozcam LOG301
Interpretation of the
Slope and the Intercept
b0 is the estimated average value of y when the
value of x is zero
ANOVA df SS MS F Significance F
450
400
350 Slope
House Price ($1000s)
300 = 0.10977
250
200
150
100
Intercept 50
= 98.248 0
0 500 1000 1500 2000 2500 3000
Square Feet
house price = 98.24833 + 0.10977 (square feet)
Dr.Burcu Ozcam LOG301
Actions When Forecasting is
Not Appropriate
• Seek information directly from customers
•Collaborate with other channel members
• Apply forecasting methods with caution (may work
where forecast accuracy is not critical)
• Delay supply response until demand
becomes clear
• Shift demand to other periods for better
supply response
• Develop quick response and flexible supply
systems
CR (2004) Prentice Hall, Inc.
Dr.Burcu Ozcam LOG301
Collaborative Forecasting
• Demand is lumpy or highly uncertain
• Involves multiple participants each with
a unique perspective—“two heads are
better than one”
• Goal is to reduce forecast error
• The forecasting process is inherently
unstable