Planning and Scheduling Operations

1
Planning and Scheduling Operations

• Forecasting

2
Forecasting

• Much has been learned in the past forty years
  about producing useful forecasts.
• 139 principles used to summarize knowledge about
  forecasting.
• Dont worry you only use some of them at any
  time. They cover
• formulating a problem, obtaining information,
  selecting applying methods, evaluating
  methods, and using forecasts.

Source Forecasting Principles, Marketing Dept,
Wharton (Business) School, University of
Pennsylvania
3
Forecasting encompasses...

• concerned with processes to make forecasts and
  assure they are properly presented and used.
• used to predict over time (time-series
  forecasting), and make predictions about
  differences among people, firms, or other objects
  (cross-sectional data).
• includes study and application of judgment and
  quantitative (statistical) methods

Source Forecasting Principles, Marketing Dept,
Wharton (Business) School, University of
Pennsylvania
4
A forecast

• a statement of the future
• is used in planning
• forecasts demand for..., AND
• skillfully blends art and science
• makes assumptions
• is rarely 100 right

Source Forecasting Principles, Marketing Dept,
Wharton (Business) School, University of
Pennsylvania
5
Forecasting and planning...

• Forecasts what future will look like,
• Planning what it should look like.
• Process usually starts with planning.
• -gt a plan!! (not surprisingly!)
• inputted to forecasting process (with information
  about environment) to create a forecast.
• Do not like the forecasts? Change the plan until
  leads to acceptable forecasts.

Source Forecasting Principles, Marketing Dept,
Wharton (Business) School, University of
Pennsylvania
6
Research into forecasting...

• Research produced many changes in recommended
  practice since 1960.
• Much advice given back then about producing
  forecasts been found to be incorrect.
• For example, developing regression models based
  upon their fit to historical data had
  detrimental effect upon forecast accuracy.

Source Forecasting Principles, Marketing Dept,
Wharton (Business) School, University of
Pennsylvania
7
Is research important to forecasting?

• Some research findings say
• relatively simple models are often more accurate
  than complex methods.

Source Forecasting Principles, Marketing Dept,
Wharton (Business) School, University of
Pennsylvania
8
Forecasting Methods Isn't common sense enough?

• quantitative methods less expensive than
  judgmental methods.
• If have to make stock control forecasts every
  week for 50,000 items, quantitative methods must
  be used.
• subjective forecasts often misleading.
• judgmental forecasts subject to many biases
  optimism and overconfidence.

Source Forecasting Principles, Marketing Dept,
Wharton (Business) School, University of
Pennsylvania
9
Forecasting Methods Isn't common sense enough?

• abysmal forecasts by experts
• "A few decades hence, energy may be free - just
  like unmetered air,"
• said by John von Neumann in1956
• IBM past forecasts?
• Others...

Source Forecasting Principles, Marketing Dept,
Wharton (Business) School, University of
Pennsylvania
10
Most commonly used methods for forecasting

• either subjective or objective.
• Subjective, or judgmental most widely used for
  important forecasts?
• Objective methods
• extrapolation (e.g., moving averages, linear
  regression against time, or exponential
  smoothing) and
• econometric methods (using regression techniques
  to estimate effects of variables).

Source Forecasting Principles, Marketing Dept,
Wharton (Business) School, University of
Pennsylvania
11
Steps in forecasting

• 1 (Setting) Purpose/objective of forecast?
• 2 Over what time period/frame?
• 3 Which forecasting technique? Select...
• 4 Collect data analyze data
• 5 Prepare forecast
• 6 Monitor/evaluate forecast
• 7 Use and present forecast

gt What variables to use gt Accuracy?
gt Structure the problem gt Who will use
forecast? gt Possible problems?
gt data sources? gt data matches? gt biased data
sources? gt diverse sources? gt valid/relevant
data? gt etc.
Source Dr. Y.-H. Chen, Production/Operations
Management, University of Texas at Dallas
12
Elements of A Good Forecast should (be)...

• have horizon with time to implement possible
  changes
• state the degree of accuracy.
• be reliable work consistently.
• be expressed in meaningful units.
• be in writing.
• be simple to understand and use, or
• be consistent with historical data intuitively.

- Timely - Accurate - Reliable -
Meaningful - Written - Easy to use
Source Dr. Y.-H. Chen, Production/Operations
Management, University of Texas at Dallas
13
Other Properties...
and..

• group(ed) forecasts tend to be more accurate
  than individual forecasts,
• errors in forecasting among a group of items
  usually cancel one another out

• accuracy decreases as time period
  increases

Source Dr. Y.-H. Chen, Production/Operations
Management, University of Texas at Dallas
14
Methods of Forecasting

• Basic Methods
• Judgmental Forecast
• Statistical (Time Series) Forecast
• Trend
• Seasonality
• Cycle
• Association

gt Averaging gt Weighted Moving
Average gt Exponential Smoothing
gt Linear Trend gt Trend-Adjusted Exponential
Smoothing
Source Dr. Y.-H. Chen, Production/Operations
Management, University of Texas at Dallas
15
Judgmental Forecast
Mostly for long-range planning and introduction
of new products. The view of one person may
prevail.

Executive opinions.
Direct customer contact composites.

• distinguish what customers would like and what
  they will actually do.
• influenced by recent sales experiences? Low sales
  lead to low estimates
• Conflict of interest. Low sales estimates lead to
  better sales performance.

Consumer survey or point-of-sales (POS) data.

• Expensive and time-consuming.
• Possible existence of irrational patterns.
• Low response rates.

Opinions of managers and staff.

• Delphi method (Rand Corp., 1948)
• etc.

Source Dr. Y.-H. Chen, Production/Operations
Management, University of Texas at Dallas
16
Delphi method

• exploration technique for forecasting, developed
  to allow discussion between experts.
• single experts sometimes suffer biases meetings
  suffer from follow the leader tendencies and
  reluctance to abandon previously stated opinions.
  
• basic notion of Delphi method is to overcome
  these shortcomings
• decision-makers have to rely on own intuition or
  on expert opinion if they lack full scientific
  knowledge.
• method been widely used to generate forecasts in
  technology, education, and other fields.
• structured process for collecting and distilling
  knowledge from group of experts, using series of
  questionnaires interspersed with controlled
  feedback.

Source THE DELPHI METHOD by research team at
Illinois Institute of Technology
17
Delphi method

• Objective(s), etc
• often derive forecasts about various aspect of
  future through collation of expert judgement.
• used to explore ideas reliably and creatively, or
  produce suitable information for decision making.
• recognises human judgement as legitimate and
  useful inputs in generating forecasts.
• useful communication device among experts and
  facilitates formation of group judgement.
• History
• development of Delphi method started in 1944.
• 1944 Theodor von Karman was asked to prepare
  future technological capabilities forecast of
  interest to military

Source THE DELPHI METHOD by research team at
Illinois Institute of Technology
18
Delphi method

• History
• 1946 Douglas Aircraft company asked to establish
  Project RAND (Research and Development) to study
  broad subject of inter-continental warfare other
  then surface.
• 1959 RAND researchers Helmer and Rescher
  published a paper which provides a philosophical
  base for forecasting.
• paper argued that in fields that have not yet
  developed to the point of having scientific laws,
  the testimony of experts is permissible.
• the problem is how to use this testimony and,
  specifically, how to combine the testimony of a
  number of experts into a single useful statement.

Source THE DELPHI METHOD by research team at
Illinois Institute of Technology
19
Statistical Data Patterns
Irregularvariation
Trend
Cycles
90
89
88
Seasonal variations
Source Dr. Y.-H. Chen, Production/Operations
Management, University of Texas at Dallas
20
Moving Averages
Source Dr. Y.-H. Chen, Production/Operations
Management, University of Texas at Dallas
21
Moving Averages
Source Dr. Y.-H. Chen, Production/Operations
Management, University of Texas at Dallas
22
Exponential Smoothing
Source Dr. Y.-H. Chen, Production/Operations
Management, University of Texas at Dallas
23
Properties of Exponential Smoothing

• Commonly used values for a range from 0.05 to
  0.50.
• Low values are used when the underlying average
  tends to be stable higher values are used when
  the underlying average is susceptible to change.
• Moving average or naive forecast can be used to
  generate starting forecast for exponential
  smoothing.

Source Dr. Y.-H. Chen, Production/Operations
Management, University of Texas at Dallas
24
Trend-Adjusted Exponential Smoothing
Source Dr. Y.-H. Chen, Production/Operations
Management, University of Texas at Dallas
25
Linear Trend
Source Dr. Y.-H. Chen, Production/Operations
Management, University of Texas at Dallas
26
Simple Seasonality Model
Source Dr. Y.-H. Chen, Production/Operations
Management, University of Texas at Dallas
27
Cycle

• Cycles are similar to seasonal variations but of
  longer duration, e.g., two to six years between
  peaks.
• It is difficult to project cycles from past data,
  because turning points are difficult to identify.
• A short moving average or a naive approach may be
  of some value.

Source Dr. Y.-H. Chen, Production/Operations
Management, University of Texas at Dallas
28
Forecast Control
It is necessary to monitor forecast errors to
ensure that the forecast is performing adequately
over time. This is generally accomplished by
comparing forecast errors to predefined values,
or action limits.
Source Dr. Y.-H. Chen, Production/Operations
Management, University of Texas at Dallas
29
Forecast Method Selection

• Most important
• Cost
• Accuracy

• Need to consider
• Historical performance
• Ability to respond to change

Source Dr. Y.-H. Chen, Production/Operations
Management, University of Texas at Dallas
30
Forecasting

• Sales forecasting research over the last four
  decades has proved one fact beyond doubt
• There is no one best forecasting model.
• You will almost always get the best forecasts by
  averaging the results from a basket of the most
  valid forecasting models.

Source Sales Forecasting for the Pharmaceutical
Industry, Inpharmation Ltd.
31
Forecasting

• The best advice is to select a few methods
  (preferably five) that seem relevant,
• make forecasts with each,
• and then take a simple average of the forecasts.

Source Forecasting Principles, Marketing Dept,
Wharton (Business) School, University of
Pennsylvania
32
Advice (contd)

• When managers receive forecasts, they often
  cannot judge their quality.
• Instead of focusing on the forecasts, however,
  they can decide whether the forecasting process
  was reasonable for the situation.
• By examining forecasting processes and improving
  them, managers may increase accuracy and reduce
  costs.

Source Standards and Practices for Forecasting,
Forecasting Principles, Marketing Dept, Wharton
(Business) School, University of Pennsylvania
33
Advice (contd)

• Forecasters often ignore common sense received
  wisdom Standards Practices for Forecasting,
  Forecasting Principles...
• Remember "Those who have knowledge, don't
  predict. Those who predict, don't have knowledge
• -- Lao Tzu, 6th Century BC Chinese Poet
• "Some things are so unexpected that no one is
  prepared for them" -- Leo Rosten in Rome Wasn't
  Burned in a Day
• "A good forecaster is not smarter than everyone
  else, he merely has his ignorance better
  organised"
• "I have seen the future and it is very much like
  the present, only longer" -- Kehlog Albran,
  The Profit

Source Famous Forecasting Quotes, D. B.
Stephenson, Reading University
34
Any Questions?

• Taken from 1. Sales Forecasting for the
  Pharmaceutical Industry, Inpharmation Ltd.,
  Henley on Thames, England http//www.inpharmati
  on.co.uk/forecasting.htm 2. Forecasting,
  Y.-H. Chen, Ph.D., Production/Operations
  Management, School of Management, University of
  Texas at Dallas 3. Forecasting Principles,
  by Professor J. Scott Armstrong, Marketing Dept,
  Wharton (Business) School, University of
  Pennsylvania http//morris.wharton.upenn.edu
  /forecast/ http//morris.wharton.upenn.edu/for
  ecast/FAQ.html 4. THE DELPHI METHOD
  written by research team led by Professor David
  Arditi at the Illinois Institute of
  Technology http//www.iit.edu/it/delphi.html
  5. Famous Forecasting Quotes, D. B.
  Stephenson, Reading University http//www.me
  t.rdg.ac.uk/cag/forecasting/quotes.html