Professional Documents
Culture Documents
Forecasting
Forecasting
xx
Forecasting Horizons
• Long Term
• 5+ years into the future
• R&D, plant location, product planning
• Principally judgement-based
• Medium Term
• 1 season to 2 years
• Aggregate planning, capacity planning, sales forecasts
• Mixture of quantitative methods and judgement
• Short Term
• 1 day to 1 year, less than 1 season
• Demand forecasting, staffing levels, purchasing, inventory levels
• Quantitative methods
Short Term Forecasting:
Needs and Uses
• Scheduling existing resources
• How many employees do we need and when?
• How much product should we make in anticipation of demand?
• Acquiring additional resources
• When are we going to run out of capacity?
• How many more people will we need?
• How large will our back-orders be?
• Determining what resources are needed
• What kind of machines will we require?
• Which services are growing in demand? declining?
• What kind of people should we be hiring?
Types of Forecasting Models
• Types of Forecasts
• Qualitative --- based on experience, judgement, knowledge;
• Quantitative --- based on data, statistics;
• Methods of Forecasting
• Naive Methods --- eye-balling the numbers;
• Formal Methods --- systematically reduce forecasting errors;
• time series models (e.g. exponential smoothing);
• causal models (e.g. regression).
• Focus here on Time Series Models
• Assumptions of Time Series Models
• There is information about the past;
• This information can be quantified in the form of data;
• The pattern of the past will continue into the future.
64 64 Pattern
Month Demand average , average
1 62 61 Average + higher of some
2 61 63 63 63 mode
3 63 65
55 Level
4 63
60
5 62
6 64 Number that
exceeds the Level + Trend
What is the forecast for period 7? 75 range + increase
63 Minimum less 1
Recession +
62.5 market cycles Cyclic
Advertisement
sales goes up Causal
Month Demand
1 62
2 61
Simple Exponential Smoothing
3 63
4 63 Ft = Dt + (1 - ) Ft-1
5 62
6 64
F0 = average = 62.5
100
1 70
90
2 78
80
3 90 70
4 103 60
5 111 0 1 2 3 4 5 6 7
6 120 Plot
130
110
100
90
80
70
60
0 1 2 3 4 5 6 7
Period Demand 130
Plot
1 70 110
2 78 100
3 90 90
4 103 80
5 111 70
6 120 60
0 1 2 3 4 5 6 7
130
Plot Y = 59.133 + 10.343t
120
Error = 2.5 70
59.133 60
0 1 2 3 4 5 6 7
Can we use the ideas from exponential smoothing?
Can we give higher weights to recent points?
Can the intercept and slope change depending on the weights?
Holt’s model
Ft+1 = at + bt.
at = α Dt + (1 – α)(at-1 + bt-1)
bt = β (at – at-1) + (1 – β)bt-1
Period Demand αD(t) (1-α)(a(t)+b(t)) a(t) β (a(t)-a(t-1)) (1-β)b(t-1) b(t) F(t+1)
1 70 14 70 10 80
2 78 15.6 64 79.6 2.88 7 9.88 89.48
3 90 18 71.584 89.584 2.9952 6.916 9.91 99.495
4 103 20.6 79.59616 100.2 3.183648 6.93784 10.1 110.32
5 111 22.2 88.2541184 110.45 3.07738752 7.08504 10.2 120.62
6 120 24 96.49323802 120.49 3.01173588 7.1137 10.1 130.62
α= 0.2
β= 0.3
Year I Year II Year III
Apr – Jun 53 58 62
Jul – Sep 22 25 27 F = 51.25
Oct – Dec 37 40 44
Jan – Mar 45 50 56
Ft+1 = (at + bt)Ct+1 where at and bt are the level and trend as described
in the Holt’s model. Ct+1 is the seasonality index for the period that
we are forecasting. The equations are
22 25 27 70
37 40 44 60
45 50 56 50
40
30
20
10
0
0 2 4 6 8 10 12 14