Beruflich Dokumente
Kultur Dokumente
Forecasting
Copyright ©2018 McGraw-Hill Higher Education. All rights reserved. No reproduction or distribution without the prior written 3-2
consent of McGraw-Hill Education
Forecast
Forecast – a statement about the future value of a
variable of interest
We make forecasts about such things as weather, demand, and
resource availability
Forecasts are important to making informed decisions
Forecasting in both environment: Design and operations
basic need for forecasting arises in estimating customer demand
for a firm’s products and services.
Need aggregate estimates of demand as well as estimates for
individual products.
long-term estimate of overall demand as well as a shorter-run
estimate of demand for each individual product or service.
Short-term demand estimates for individual products are
necessary to determine daily or weekly management of the firm’s
activities such as scheduling personnel and ordering materials.
Copyright ©2018 McGraw-Hill Higher Education. All rights reserved. No reproduction or distribution without the prior 3-3
written consent of McGraw-Hill Education
Two Important Aspects of Forecasts
Expected level of demand
The level of demand may be a function of some structural
variation such as trend or seasonal variation
Accuracy
Related to the potential size of forecast error
Copyright ©2018 McGraw-Hill Higher Education. All rights reserved. No reproduction or distribution without the prior 3-4
written consent of McGraw-Hill Education
Forecast Uses
the difference between forecasting and planning.
Planning is often in response to a forecast. A passive response
would be to reduce output because of a predicted decrease in
demand, while an active response would be to advertise in an
effort to offset the predicted decrease in demand.
Plan the system
Generally involves long-range plans related to:
Types of products and services to offer
Facility and equipment levels
Facility location
Plan the use of the system
Generally involves short- and medium-range plans related to:
Inventory management
Workforce levels
Purchasing
Production
Budgeting
Scheduling
Copyright ©2018 McGraw-Hill Higher Education. All rights reserved. No reproduction or distribution without the prior written 3-5
consent of McGraw-Hill Education
Features Common to All Forecasts
1. Techniques assume some underlying causal system that existed
in the past will persist into the future
2. Forecasts are not perfect
3. Forecasts for groups of items are more accurate than those for
individual items
4. Forecast accuracy decreases as the forecasting horizon increases
Copyright ©2018 McGraw-Hill Higher Education. All rights reserved. No reproduction or distribution without the 3-6
LO 3.1 prior written consent of McGraw-Hill Education
Forecasts Are Not Perfect
Forecasts are not perfect:
Because random variation is always present, there will always
be some residual error, even if all other factors have been
accounted for.
Customers themselves may be unclear regarding what
they'd like to do versus what they'll actually do.
Which of the following is a potential shortcoming of using sales force opinions in
demand forecasting?
A. Members of the sales force often have substantial histories of working with and
understanding their customers.
B. Members of the sales force often are well aware of customers' future plans.
C. Members of the sales force have direct contact with consumers.
D. Members of the sales force can have difficulty distinguishing between what
customers would like to do and what they actually will do.
E. Customers often are quite open with members of the sales force with regard to
future plans. 3-7
Copyright ©2018 McGraw-Hill Higher Education. All rights reserved. No reproduction or distribution without
LO 3.2 the prior written consent of McGraw-Hill Education
Elements of a Good Forecast
The forecast
Should be timely
Should be accurate (enable users to plan for possible errors )
Should be reliable
Should be expressed in meaningful units
Should be in writing
Technique should be simple to understand and use
Should be cost-effective (The benefits should outweigh the
costs)
Copyright ©2018 McGraw-Hill Higher Education. All rights reserved. No reproduction or distribution without 3-8
LO 3.3 the prior written consent of McGraw-Hill Education
Six Steps in the Forecasting Process
1. Determine the purpose of the forecast. This step will provide an
indication of the level of detail required in the forecast, the amount
of resources that can be justified, and the level of accuracy
necessary.
2. Establish a time horizon. The forecast must indicate a time interval,
keeping in mind that accuracy decreases as the time horizon increases.
3. Obtain, clean, and analyze appropriate data. Once obtained,
the data may need to be “cleaned” to get rid of outliers and
obviously incorrect data before analysis.
4. Select a forecasting technique.
5. Make the forecast.
6. Monitor the forecast errors. Monitor the forecast performance
and reexamine the method, assumptions, validity of data, and so
on; modify as needed; and prepare a revised forecast.
Copyright ©2018 McGraw-Hill Higher Education. All rights reserved. No reproduction or distribution without 3-9
LO 3.4 the prior written consent of McGraw-Hill Education
Forecast Accuracy and Control
Allowances should be made for forecast errors
It is important to provide an indication of the extent to
which the forecast might deviate from the value of the
variable that actually occurs
Forecast errors should be monitored
Error = Actual – Forecast
If errors fall beyond acceptable bounds, corrective action
may be necessary
Copyright ©2018 McGraw-Hill Higher Education. All rights reserved. No reproduction or distribution without 3-10
LO 3.5 the prior written consent of McGraw-Hill Education
Forecast Accuracy Metrics
MAD
Actual t Forecast t Mean Absolute Deviation
n (MAD) weights all errors evenly
Actual t Forecast t 2
Mean Square Error (MSE)
MSE
n 1 weights errors according to their
squared values
Actualt Forecast t
Actual t
100
Mean Absolute Percent Error
MAPE
n (MAPE) weights errors
according to relative error
Copyright ©2018 McGraw-Hill Higher Education. All rights reserved. No reproduction or distribution without the 3-11
LO 3.5 prior written consent of McGraw-Hill Education
Forecast Error Calculation
Actual Forecast (A-F)
Period
(A) (F) Error |Error| Error2 [|Error|/Actual]x100
1 107 110 -3 3 9 2.80%
Sum 13 39 11.23%
Copyright ©2018 McGraw-Hill Higher Education. All rights reserved. No reproduction or distribution without the 3-12
LO 3.5 prior written consent of McGraw-Hill Education
Forecasting Approaches
1-Qualitative forecasting (Judgmental forecasts)
Qualitative techniques permit the inclusion of soft information such as:
Human factors
Personal opinions
Hunches
These factors are difficult, or impossible, to quantify
2- Quantitative forecasting
These techniques rely on hard data
Quantitative techniques involve either the projection of historical data
or the development of associative methods that attempt to use causal
variables to make a forecast
use equations that consist of one or more explanatory variables that can
be used to predict demand. For example, demand for paint might be
related to variables such as the price per gallon and the amount spent
on advertising, as well as to specific characteristics of the paint (e.g.,
drying time, ease of cleanup).
Copyright ©2018 McGraw-Hill Higher Education. All rights reserved. No reproduction or distribution without 3-13
LO 3.6 the prior written consent of McGraw-Hill Education
Qualitative Forecasts
Forecasts that use subjective inputs such as opinions from
consumer surveys, sales staff, managers, executives, and experts
Executive opinions
A small group of upper-level managers may meet and collectively develop a
forecast
Sales force opinions
Members of the sales or customer service staff can be good sources of
information due to their direct contact with customers and may be aware of
plans customers may be considering for the future
Consumer surveys
Since consumers ultimately determine demand, it makes sense to solicit input
from them
Consumer surveys typically represent a sample of consumer opinions
Other approaches
Managers may solicit 0pinions from other managers or staff people or outside
experts to help with developing a forecast.
The Delphi method is an iterative process intended to achieve a consensus
Copyright ©2018 McGraw-Hill Higher Education. All rights reserved. No reproduction or distribution without 3-14
LO 3.6 the prior written consent of McGraw-Hill Education
Forecasting Approaches
3- Time-Series Forecasts
Forecasts that project patterns identified in recent
time-series observations
Time-series – a time-ordered sequence of observations
taken at regular time intervals
Assume that future values of the time-series can be
estimated from past values of the time-series
The data may be measurements of demand, earnings,
profits, shipments, accidents, output, precipitation,
productivity, or the consumer price index.
Can often be accomplished with a plot (Trend, seasonal
variations…etc)
Copyright ©2018 McGraw-Hill Higher Education. All rights reserved. No reproduction or distribution without the prior 3-15
written consent of McGraw-Hill Education
Time-Series Behaviors
1. Trend refers to a long-term upward
or downward movement in the data.
2. Seasonality refers to short-term,
fairly regular variations generally
related to factors such as the
calendar or time of day.
3. Cycles are wavelike variations of
more than one year’s duration.
4. Irregular variations are due to
unusual circumstances such as
severe weather conditions, strikes,
or a major change in a product or
service. A demand forecast should be based
on a time series of past demand
5. Random variations are residual rather than unit sales. Sales would
variations that remain after all other not truly reflect demand if one or
behaviors
Copyright haveHigher
©2018 McGraw-Hill been accounted
Education. for.
All rights reserved. moreorstockouts
No reproduction occurred.
distribution without the prior written 3-16
consent of McGraw-Hill Education
Examples of Time series forecasting
Trend
Population shifts
Changing income
Seasonality
Restaurants, service call centers, and theaters all experience seasonal demand
Cycle
Wavelike variations lasting more than one year
These are often related to a variety of economic, political, or even agricultural
conditions
Irregular variation
Due to unusual circumstances that do not reflect typical behavior
Labor strike
Weather event
Random Variation
Residual variation that remains after all other behaviors have been accounted
for
Copyright ©2018 McGraw-Hill Higher Education. All rights reserved. No reproduction or distribution without the prior written 3-17
consent of McGraw-Hill Education
Time-Series Forecasting - Naïve Forecast
Naïve forecast
Uses a single previous value of a time series as the basis for a forecast
The forecast for a time period is equal to the previous time
period’s value
Can be used with
A stable time series
Seasonal variations
Trend
For data with trend, the forecast is equal to the last value of the series
plus or minus the difference between the last two values of the series. For
example, suppose the last two values were 50 and 53. The next forecast
would be 56:
Copyright ©2018 McGraw-Hill Higher Education. All rights reserved. No reproduction or distribution without 3-18
LO 3.7 the prior written consent of McGraw-Hill Education
Time-Series Forecasting - Averaging
These techniques work best when a series tends to vary
about an average
Averaging techniques smooth variations in the data
They can handle step changes or gradual changes in the level
of a series
Techniques
1. Moving average
2. Weighted moving average
3. Exponential smoothing
Moving Average. One weakness of the naive method is that the forecast just traces the actual
data, with a lag of one period; it does not smooth at all. But by expanding the amount of
historical data a forecast is based on, this difficulty can be overcome. A moving average
forecast uses a number of the most recent actual data values in generating a forecast.
Copyright ©2018 McGraw-Hill Higher Education. All rights reserved. No reproduction or distribution without 3-19
LO 3.8 the prior written consent of McGraw-Hill Education
Moving Average
Technique that averages a number of the most recent
actual values in generating a forecast.
Example:
Compute a three-period moving average
forecast given demand for shopping carts
for the last five periods.
Copyright ©2018 McGraw-Hill Higher Education. All rights reserved. No reproduction or distribution without 3-22
LO 3.9 the prior written consent of McGraw-Hill Education
Exponential Smoothing
A weighted averaging method that is based on the previous
forecast plus a percentage of the forecast error ( α )
Ft Ft 1 ( At 1 Ft 1 ) For example, suppose the previous
forecast was 42 units, actual demand
where was 40 units, and α = .10. The new
Ft Forecast for period t forecast would be computed as
Ft 1 Forecast for the previous period follows:
Ft=42+.10(40-42)=41.8
= Smoothing constant
At 1 Actual demand or sales from the previous period Then, if the actual demand turns out
to be 43, the next forecast would be:
Ft=41.8+.10(43-41.8)=41.92
The quickness of forecast adjustment to error is determined by the smoothing constant, α. The
closer its value is to zero, the slower the forecast will be to adjust to forecast errors (i.e., the
greater the smoothing).
Selecting a smoothing constant is basically a matter of judgment or trial and error, using
forecast errors to guide the decision. The goal is to select a smoothing constant that balances
the benefits of smoothing
Copyright randomHigher
©2018 McGraw-Hill variations
Education.with thereserved.
All rights benefits of responding
No reproduction to real
or distribution changes
without if
3-23
LO 3.10 the prior written consent of McGraw-Hill Education
and when they occur.
Exercise: Compare the error performance of these three forecasting techniques
using MAD, MSE, and MAPE: a naive forecast, a two-period moving average, and
exponential smoothing with α = .10 for periods 3 through 11, using the data shown
Fill in the following table and calculate the MAD,MSE and MAPE
3-24
Linear Trend
A simple data plot can reveal the existence and nature of a
trend Ft a bt
where
Linear trend equation
Ft Forecast for period t
For example, consider the trend a Value of Ft at t 0
equation Ft = 45 + 5t. The value b Slope of the line
of Ft when t = 0 is 45, and the t Specified number of time periods from t 0
slope of the line is 5, which
means that, on the average, the
value of Ft will increase by five
units for each time period.
If t = 10, the forecast, Ft, is:
45 + 5(10) = 95 units.
Copyright ©2018 McGraw-Hill Higher Education. All rights reserved. No reproduction or distribution without the 3-25
LO 3.11 prior written consent of McGraw-Hill Education
Estimating Slope and Intercept
Cell phone sales for a California-based firm
Slope and intercept over the last 10 weeks are shown in the
can be estimated following table. Plot the data, and visually
from historical data check to see if a linear trend line would be
appropriate. Then determine the equation of
n ty t y the trend line, and predict sales for weeks 11
b
n t t and 12.
2
2
a
y b t
or y bt
n
where
n Number of periods
y Value of the time series
b = 7.51 and a = 699.40 The trend line is Ft = 699.40 + 7.51t,
Substituting values of t into this equation, the forecasts for the next two
periods (i.e., t = 11 and t = 12) are:
1. F11 = 699.40 + 7.51(11) = 782.01 3-26
Copyright ©2018 McGraw-Hill Higher Education. All rights reserved. No reproduction or distribution without the
LO 3.11
2. Fconsent
prior written
12 = 699.40 + 7.51(12) = 789.52
of McGraw-Hill Education
t Y t*Y t2
1 405 405 1
2 410 820 4
3 420 1,260 9
4 415 1,660 16
5 412 2,060 25
6 420 2,520 36
7 424 2,968 49
8 433 3,464 64
9 438 3,942 81
10 440 4,400 100
11 446 4,906 121
12 451 5,412 144
13 455 5,915 169
14 464 6,496 196
15 466 6,990 225
16 474 7,584 256
17 476 8,092 289
18 482 8,676 324
171 7,931 77,750 2,109
n tY t Y 18(77,750) 171(7,931)
b 4.59
n t 2 ( t ) 2 18(2,109) (171) 2
Y b t 7,931 4.59(171)
a 397.01
n 18
Y19 = 397.01 + (4.59)(19) = 484.22
Y20 = 397.01 + (4.59)(20) = 488.81
Y21 = 397.01 + (4.59)(21) = 493.40
Starting forecast =
728 + 9.33 = 737.33
Unlike a linear trend line,
trend-adjusted smoothing has
the ability to adjust to
changes in trend.
Copyright ©2018 McGraw-Hill Higher Education. All rights reserved. No reproduction or distribution without 3-29
LO 3.12 the prior written consent of McGraw-Hill Education
The initial estimate of trend is based on
the net change of 30 for the three
periods from 1 to 4, for an average of
+10 units. Use = .5 and = .4.
Round values to two decimals.
Initial trend = (240 – 210)/3 = 10.00
3-30
Techniques for Seasonality Knowledge of seasonal
variations is an important
Seasonality – regularly repeating factor in retail planning and
movements in series values that can scheduling. Moreover,
seasonality can be an
be tied to recurring events important factor in capacity
Expressed in terms of the amount that
planning for systems that
actual values deviate from the average must be designed to handle
value of a series peak loads
Models of seasonality
Additive
Seasonality is expressed as a quantity
that gets added to or subtracted from the
time-series average in order to
incorporate seasonality
Multiplicative
Seasonality is expressed as a percentage
of the average (or trend) amount which
is then used to multiply the value of a
series in order to incorporate seasonality
Copyright ©2018 McGraw-Hill Higher Education. All rights reserved. No reproduction or distribution without 3-31
LO 3.13 the prior written consent of McGraw-Hill Education
Suppose that the seasonal relative for the
Seasonal Relatives quantity of toys sold in May at a store is
Seasonal relatives: The seasonal 1.20. This indicates that toy sales for that
percentage used in the multiplicative month are 20 percent above the monthly
seasonally adjusted forecasting average.
model A seasonal relative of .90 for July indicates
Using seasonal relatives that July sales are 90 percent of the
monthly average.
To deseasonalize data
Done in order to get a clearer A coffee shop owner wants to estimate
picture of the nonseasonal (e.g., demand for the next two quarters for hot
trend) components of the data chocolate. Sales data consist of trend
series and seasonality.
Divide each data point by its 1.Quarter relatives are 1.20 for the first
seasonal relative quarter, 1.10 for the second quarter,
To incorporate seasonality in a 0.75 for the third quarter, and 0.95 for
b) The forecast
trend values are: the fourth quarter. Use this
Period 9:1.Ft =Obtain
124 + 7.5(9) trend = 191.5
estimates for information to deseasonalize sales for
Period 10: Ft desired
= 124 + 7.5(10)periods using a trend
= 199.0 quarters 1 through 8.
equation
Period 9 is a first quarter and period 10 is a 2.Using the appropriate values of
2. Add
second quarter. seasonality
Multiplying eachby multiplying
trend value by quarter relatives and the equation Ft =
the appropriatethese
quarter trend estimates
relative results in: by the
corresponding seasonal relative 124 + 7.5t for the trend component, 3-32
Period 9:
LO 3.13 191.5(1.20)
Copyright =
©2018 229.8
McGraw-Hill Higher Education. All rights reserved. No reproduction or distribution without the
prior written consent of McGraw-Hill Education
Period 10: 199.0(1.10) = 218.9 estimate demand for periods 9 and 10.
Computing Seasonal Relatives Using the Simple Average Method. The simple
average (SA) method is an alternative way to compute seasonal relatives. Each seasonal
relative is the average for that season divided by the average of all seasons. This method is
illustrated in Example 8B, where the seasons are days. Note that there is no need to
standardize the relatives when using the SA method.
3-33
Monitoring the Forecast
Tracking forecast errors and analyzing them can provide useful insight
into whether forecasts are performing satisfactorily
Sources of forecast errors:
The model may be inadequate due to
a. omission of an important variable
b. a change or shift in the variable the model cannot handle
c. the appearance of a new variable
Irregular variations may have occurred
Random variation
Control charts are useful for identifying the presence of non-random
error in forecasts
Tracking signals can be used to detect forecast bias
LO 3.15 Copyright ©2018 McGraw-Hill Higher Education. All rights reserved. No reproduction or distribution 3-34
without the prior written consent of McGraw-Hill Education
Control Chart Construction
1. Compute the MSE.
2. Estimate of standard deviation of the distributi on of errors
s MSE
3. UCL : 0 z MSE
4. LCL : 0 z MSE
where z Number of standard deviations from the mean
To construct a control chart, first compute the MSE. The square root
of MSE is used in practice as an estimate of the standard deviation
of the distribution of errors.
3-35
LO 3.15 Copyright ©2018 McGraw-Hill Higher Education. All rights reserved. No reproduction or
distribution without the prior written consent of McGraw-Hill Education
Choosing a Forecasting Technique
Factors to consider Factors to consider
Cost Availability of forecasting software
Accuracy Time needed to gather and analyze
Availability of data and prepare a forecast
historical data Forecast horizon
Copyright ©2018 McGraw-Hill Higher Education. All rights reserved. No reproduction or distribution without 3-36
LO 3.16 the prior written consent of McGraw-Hill Education
Operations Strategy
The better forecasts are, the more able organizations will
be to take advantage of future opportunities and reduce
potential risks
A worthwhile strategy is to work to improve short-term forecasts
Accurate up-to-date information can have a significant effect on
forecast accuracy:
Prices
Demand
Other important variables
Reduce the time horizon forecasts have to cover
Sharing forecasts or demand data through the supply chain can
improve forecast quality
Copyright ©2018 McGraw-Hill Higher Education. All rights reserved. No reproduction or distribution without the 3-37
prior written consent of McGraw-Hill Education