• Introduction • Reasons • Forecasting Methods
• Introduction• Reasons• Forecasting Methods
Introduction
Forecasting : Estimating the future demand.
Demand estimates for products and services are the starting point for all the other planning in operations management.
Management teams develop sales forecasts based in part on demand estimates.
The sales forecasts become inputs to both business strategy and production resource forecasts.
Why Forecasting is Essential in OM
New Facility Planning – It can take 5 years to design and build a new factory or design and implement a new production process. Therefore long-term demand forecast is necessary.
Production Planning – Demand for products vary from month to month and it can take several months to change the capacities of production processes. Therefore medium-range forecast is done.
Workforce Scheduling – Demand for services (and the necessary staffing) can vary from hour to hour and employees weekly work schedules must be developed in advance. Here short-term forecasting is needed.
Forecasting Methods
Qualitative Method
○ Used when situation is vague & small data exist◦ New products
◦ New technology
○ Involves intuition, experience
Quantitative Method
○ Used when situation is ‘stable’ & historical data exist◦ Existing products
◦ Current technology
○ Involves mathematical techniques
Qualitative Methods • Grass Roots
– Person closed to the customer, knows the need of customer better.
– Forecasting is done by compiling input from those at the end of the
hierarchy who deal with what is being forecasted.
• Market research– Used mostly for product research and for new product sales.
– Data is collected in a variety of ways : Surveys and interviews etc.
• Panel consensus– The idea is the group will produce better forecasts than an individual.
• Historical analogy– Important in planning new products where a forecast may be derived by
using the history of a similar product.
• Delphi method– Group of experts responds to questionnaire.
– Moderator compiles results and formulates a new questionnaire which is submitted to the group.
Delphi method
1. Choose the experts to participate representing a variety of knowledgeable people in different areas
2. Through a questionnaire (or E-mail), obtain forecasts (and any premises or qualifications for the forecasts) from all participants
3. Summarize the results and redistribute them to the participants along with appropriate new questions
4. Summarize again, refining forecasts and conditions, and again develop new questions
5. Repeat Step 4 as necessary and distribute the final results to all participants
Quantitative Method
Time series AnalysisA time series is a set of numbers where the order or
sequence of the numbers is important, e.g., historical
demand
Analysis of the time series identifies patterns
Once the patterns are identified, they can be used to
develop a forecast
Components of a Time Series
VariationsCyclic Variation is a data pattern that may cover several years
before it repeats itself.Seasonal Variation is a data pattern that repeats itself over the
period of one year or less.Random Variation results from random variation or unexplained
causes.
Trends are noted by an upward or downward sloping line.
“Measurement of Trend” & “Demand Forecasting” Graphical Method
Method of Semi-Averages
Method of least squares (Long range method)
Data Set
Fit a trend line and find whether sales are decreasing or increasing.Year Annual sales (000’s) Semi-average1991 531992 791993 76 437/6 = 72.83 (000’s) 1994
661995 691996 941997 1051998 871999 792000 104 560/6 = 93.33 (000’s)2001 972002 922003 101
Graphical Method
0
20
40
60
80
100
120
1990 1995 2000 2005
years
Sal
es (
000'
s)
Semi-average Method (n-odd)
0
20
40
60
80
100
120
1990 1995 2000 2005
years
Sal
es (
000'
s)
Semi-average Method (n-even)
Fit a trend line and find whether sales are decreasing or increasing.Year Annual sales (000’s) Semi-average1991 531992 791993 76 343/5 = 68.61994 661995 691996 941997 1051998 87 469/5 = 93.81999 792000 104
Semi-average Method (n-even)
0
20
40
60
80
100
120
1990 1992 1994 1996 1998 2000 2002
Years
An
nu
al S
les
Least Square Method (n-odd)(Long Range Method)
In a production unit, the production of a handles during the years 1993-2003 is given in the following data :Year(X) Production (000’s) t t.Ut Trend
values(Ut)
1993 66.6 -5 -333.0 25 75.741994 84.6 -4 -339.6 16 79.691995 88.6 -3 -265.8 9 83.64 1996 78.0 -2 -156.0 4 87.591997 96.8 -1 -96.8 1 91.541998 105.2 0 0 0 95.491999 93.2 1 93.2 1 99.442000 111.6 2 223.2 4 103.392001 88.3 3 264.9 9 107.342002 117.0 4 468.0 16 111.292003 115.2 5 576.0 25 115.24
2t
Least Square Method (n-odd)
0
20
40
60
80
100
120
140
1992 1994 1996 1998 2000 2002 2004
years
Pro
du
ctio
n
Production (000’s) Trend Values
Least Square Method
t=X-1998 (n-odd) Ut=a+bt (straight line) ∑Ut=na+b∑t (1) ∑t.Ut=a∑t+b∑t.t (2) When ∑t=0, a=∑Ut/n
b=∑t.Ut/ ∑ t=X-[(1997+1998)/2] (n-even)
(interval)/2
2t
Least-Square Method (n-even)
Forecast the sales of the bags for next two years and tell the movement of the trend.Years Sales (lakhs)2000 762001 802002 1302003 1442004 1382005 1202006 1742007 190
Least-Square Method (n-even)
Forecast the sales of the bags for next two years and tell the movement of the trend.Years Sales t t.Ut Trend values (X) (lakhs)(Ut) Ut=131.5+7.33t2000 76 -7 -532 49 80.192001 80 -5 -400 25 94.852002 130 -3 -390 9 109.512003 144 -1 -144 1 124.172004 138 1 138 1 138.832005 120 3 360 9 153.492006 174 5 870 25 168.152007 190 7 1330 49 182.18
Total ∑Ut=1052 ∑t=0 ∑t.Ut=1232 ∑ =168
Sales for year 2008 = Ut=131.5+7.33*(9) = 197.47Sales for year 2009 = Ut=131.5+7.33*(11) = 212.13
2t
2t
Least-Square Method (n-even)
0
50
100
150
200
1998 2000 2002 2004 2006 2008
Years
Sal
es
Sales Trend values
Simple Moving Average (Demand Forecasting for Short Range)
The simple moving average model assumes an average is a good estimator of future behavior
The formula for the simple moving average is:
Ft = Forecast for the coming periodn = Number of periods to be averagedAt-1 = Actual occurrence in the past period for up to “n” periods
n
AAAAF ntttt
t
.....321
Simple Moving Average Week Demand 3-Week 6-Week
1 650
2 678
3 720
4 785 682.67
5 859 727.67
6 920 788.00
7 850 854.67 768.67
8 758 876.33 802.00
9 892 842.67 815.33
10 920 833.33 844.00
11 789 856.67 866.50
12 844 867.00 854.83
F4=(650+678+720)/3
F7=(650+678+720+785+859
+920)/6
Demand for week 13 (for 3-week moving-average) = (920+789+844)/3 = 851
Demand for week 13 (for 6-week) = (850+758+892+920+789+844)/3 = 842.17
Simple Moving Average
0
200
400
600
800
1000
0 2 4 6 8 10 12 14
Weeks
Dem
and
Demand 3-Week 6-Week
Simple Moving Average
It is called “moving” because as new demand data becomes available, the oldest data is not used
By increasing the Averaging Period, the forecast is less responsive to fluctuations in demand
By decreasing the Averaging Period, the forecast is more responsive to fluctuations in demand
Exponential Smoothing(Short – Range Forecast)
In the exponential smoothing method, three pieces of data are needed to forecast the future : the most recent forecast, the actual demand and a smoothing constant alpha (α).
The smoothing constant (α) determines the level of smoothing and speed of reactions for that product.
Smoothing constant (α) be given a value between 0 and 1.If the real demand is stable (such as demand for electricity or
food), then small α should be chosen.If the real demand is rapidly increasing or decreasing (such as
fashion items), then large α is chosen.
111 tttt FAFF
Exponential Smoothing Method
In the formula : forecast for period t, the next period
forecast for period t-1, the prior period
actual data for period t-1, the prior period
Initially consider :Forecasted demand for prior period Ft-1 = Actual demand of current period At
Actual demand for prior period At-1 = Actual demand of current period At
1tF
1tA
tF
Example : A company talks to an analyst at company headquarters about forecasting weekly demand for inventory from company’s warehouse. The analyst suggests that company consider using exponential smoothing with smoothing constant ‘α’ of 0.1, 0.2, and 0.3. Company decides to compare the accuracy of ‘α’ .
WeekActual inventory
demandForecast at
α = 0.1Forecast at
α = 0.2Forecast at
α = 0.3
1 85 85 85 85
2 102 85 85 85
3 110 86.7 88.4 90.1
4 90 89.0 92.7 96.0
5 105 89.1 92.1 94.2
6 95 90.7 94.7 97.4
7 115 91.1 94.7 96.7
8 120 93.5 98.8 102.2
8 80 96.1 103.1 107.5
9 95 94.5 98.4 99.2
10 100 94.6 97.7 97.9
Comparison w.r.t. smoothing constant
α = 0.1 α = 0.2 α = 0.3
Week
Actual inventory demand Forecast
Absolute Deviation Forecast
Absolute Deviation Forecast
Absolute Deviation
1 85 85 0 85 0 85 0
2 102 85 17 85 17 85 17
3 110 86.7 23.3 88.4 21.6 90.1 19.9
4 90 89.0 0.9 92.7 2.7 96.0 6.0
5 105 89.1 15.8 92.2 12.8 94.2 10.7
6 95 90.7 4.2 94.7 0.2 97.4 2.4
7 115 91.1 23.8 94.8 20.2 96.7 18.2
8 120 93.5 26.4 98.8 21.1 102.2 17.7
9 80 96.2 16.1 103.1 23.0 107.5 27.5
10 95 94.5 0.4 98.4 3.4 99.2 4.2
11 100 94.6 5.4 97.7 2.2 97.9 2.0 α = 0.2 gives slightly better accuracy than α =0.1 & α = 0.3
Exponential Smoothing Method
0
20
40
60
80
100
120
140
0 2 4 6 8 10 12
Weeks
Dem
and
Actual inventory demand α = 0.1 α = 0.2 α = 0.3
Example A toy company buys large quantities of plastic pellets for use in the manufacture of
its products. Production manager wants to develop a forecasting system for plastic
pellet prices. The price per pound of plastic pellets has varied as shown:
Plastic Pellets Plastic Pellet
Month Price/Pound Month Price/Pound
1 0.39 11 0.39
2 0.41 12 0.43
3 0.45 13 0.37
4 0.44 14 0.38
5 0.41 15 0.36
6 0.41 16 0.39
7 0.38 Use exponential smoothing to forecast monthly
8 0.36 Plastic pellet prices for alpha as 0.1, 0.3, 0.5
9 0.35 and justify which alpha value is giving good
10 0.38 results.