Value Focused Thinking Systems 473

1
Value Focused ThinkingSystems 473
Based on Slides by Dr A. Loerch
2
Where are we going?

• Our goal is to build a Value Function
• Quantifies the benefit that we get from
  implementing a decision alternative
• Type of utility function
• Typically maps anticipated benefit to a
  dimensionless number (utility)
• Used for comparison purposes (which alternatives
  do more to give the decision make what he/she
  wants)
• Can be used as an objective function for
  optimization
• Helps overcome the things that make a decision
  hard to make
• First step in building a value function is to
  identify the Decision-makers (DMs) values

3
Overall Method We Will Employ

• Start with Objective
• Define values (not alternatives)
• Generate better alternatives based on values
• Evaluate alternatives using values
• This is the Value Focused Thinking Philosophy
• Multiple Objective Decision Analysis is the
  mathematical technique used to implement VFT
• Also called
• Multiple Attribute Utility Theory
• Multiple Attribute Value Theory
• Multiple Attribute Preference Theory

4
Building a Qualitative Value Model

• Step 1 Identify the fundamental objective
• Step 2 Identify functions that provide value
• Step 3 Identify the objectives that define value
  
• Step 4 Identify the value measures

5
Values

• Values Those things that are important to a
  decision maker, things that they care about
• Values are fundamental to all we do
• Values should drive decisions
• Alternative-Focused Thinking
• Standard Approach
• Problem arises
• Problem solving begins
• Alternative solutions listed
• Best alternative chosen
• Limited way to think
• Reactive not proactive
• Incomplete analysis
• Backward thinking

6
Identifying Values

• Interact with decision makers
• Elicit feedback throughout design process or
  analysis
• Values are stated through objectives
• 3 Features of objectives
• (1) Decision Context
• (2) Object
• (3) Direction of Preference
• ex. Forest products company wants to minimize
  environmental impact
• (1) Decision Context - Harvesting Trees
• (2) Object - Environmental Impact
• (3) Direction of Preference - Minimum is best

7
Where does thinking about values lead?
Keeney, Ralph L., Value-Focused Thinking
A Path To Creative Decisionmaking, Harvard
University Press, Cambridge, MA, 1992, pp. 3-28.
8
Ways to Identify Valuesdue to Professor Gregory
Parnell, USMA

• Gold Standard based on an approved vision,
  policy, strategy, planning, or doctrine document
• Values have been thought about, discussed, and
  written down
• We use work that has already been done and
  approved
• Platinum Standard based on interviews with
  decision-makers and stakeholders
• Often difficult to get enough time with
  stakeholders and DMs
• Use Affinity Diagram for group settings (will
  discuss later)
• Silver Standard uses data provided by
  stakeholder representatives
• When real DMs and stakeholders are not available
• Still use Affinity Diagrams for groups
• Combined Standard combination of the above

9
The University's Mission

• George Mason University will be an institution of
  international academic reputation providing
  superior education for students to develop
  critical, analytical, and imaginative thinking
  and to make well-founded ethical decisions. It
  will respond to the call for interdisciplinary
  research and teaching...
• The university will prepare students to address
  the complex issues facing them in society and to
  discover meaning in their own lives. It will
  encourage diversity in its student body and will
  meet the needs of students by providing them with
  interdisciplinary and innovative undergraduate,
  graduate, and professional courses. The
  university will energetically seek ways to
  interact with and serve the needs of the student
  body.
• The university will nurture and support a faculty
  that is diverse, innovative, excellent in
  teaching, active in pure and applied research,
  and responsive to the needs of students and the
  community. The faculty will embody the
  university's interactive approach to change both
  in the academy and in the world.
• The university will be a resource of the
  Commonwealth of Virginia serving private and
  public sectors. It will be an intellectual and
  cultural nexus between Northern Virginia, the
  nation, and the world.

10
Process of Identifying Values

• Research potential Gold Standard documents
• Recognize that environment might have changed
• Results of research must be validated with DMs
• Interview Stakeholders and/or Decision Makers
• Validate Gold Standard work
• If none, elicit values directly
• Interact with lower level representatives as
  necessary
• Fill in the blanks
• Make sure you are consistent with thoughts of DMs
• Combine above as necessary
• Note always check regularly with the decision
  makers or reps!

11
Structuring Objectives

• Initial list has non-objectives
• Alternatives
• Constraints
• Criteria for evaluation
• Convert non-objectives into objectives
• Means objectives
• Fundamental objectives
• Why is each important?
• It leads to something else Means Objective
• It is essential Fundamental Objective

12
Values help guide strategic thinking
Fundamental
Objective
Means Obj 1
Means Obj 2
Means Obj 3

• Strategic objectives should guide our
  decision-making
• Strategic objectives are stable over time
  although the
• means to obtain these objectives vary over time.

13
My professional values help guide my strategic
planning
Fundamental Objective
Means Objectives
14
Techniques or Eliciting Values

• (1) Make a Wish List
• (2) Evaluation of alternatives
• (3) Problems and Shortcomings
• (4) Consequences or Outcomes
• (5) Goals, Constraints, Guidelines
• (6) Different perspectives
• (7) Strategic Objectives
• (8) Generic Objectives
• (9) Structure of Objectives
• (10) Quantifying Objectives

15
Constructing Value Hierarchies
16
Value Hierarchy Exercise

• You want to build / buy a new computer
• You are the decision maker and analyst
• Overall Value A good computer
• Task construct a hierarchy of values
• Concepts not numbers
• Qualifiers
• Later Convert that hierarchy to an objective
  hierarchy
• Decision Context (value)
• Measure
• Direction of preference

17
Ways to Identify Valuesdue to Professor Gregory
Parnell, USMA

• Gold Standard based on an approved vision,
  policy, strategy, planning, or doctrine document
• Values have been thought about, discussed, and
  written down
• We use work that has already been done and
  approved
• Platinum Standard based on interviews with
  decision-makers and stakeholders
• Often difficult to get enough time with
  stakeholders and DMs
• Use Affinity Diagram for group settings (will
  discuss later)
• Silver Standard uses data provided by
  stakeholder representatives
• When real DMs and stakeholders are not available
• Still use Affinity Diagrams for groups
• Combined Standard combination of the above

18
Types of Measures

• Natural In general use and have a common
  interpretation to everyone.
• Constructed Integrates multiple numerical and/
  or verbal descriptions into a single description
  of the state of a fundamental objective. Applies
  to a specific decision context.
• Proxy Normally means objectives that have a
  derived, or implied relationship to a fundamental
  objective. Substitutes for a desired measure

19
Natural Measures

• Speed in Miles per hour
• Cost in U.S. Dollars
• Weight in Tons
• Size in cubic feet

20
Constructed Measures

• Usually a collection of factors
• Context Dependent
• Examples
• 1, 2,?, 10, where 1 is best and 10 is worse
• 1, 0, 1, where 0 indicates the status quo,
• 1 is something worse, and 1 is
  something better

21
Proxy Measures

• A proxy measure is something that we can measure
  which replaces an objective or decision criterion
  that cant be measured, is hard or expensive to
  measure, would require subjective measures, etc.
• Used because it relates to a fundamental
  objective.
• Could be a natural measure, such as pounds
• Example Vehicle weight in pounds as a proxy for
  smooth driving ? Heavier cars tend to ride better.

22
Why Do We Care About Measures?

• When we have multiple objectives (common in more
  realistic decision problems), each decision
  criterion associated with an objective will be
  measured using a natural, constructed or proxy
  measure
• The book uses the term attributes instead of
  measures
• The multiple measures will normally be converted
  to value or utility functions (more on that
  later), which have the same scale (usually 0 to
  1)
• The type of measure will often suggest the
  correct form of a value or utility function
  e.g., linear, exponential convex, exponential
  concave
• If you dont get the measures right, then the
  entire decision analysis will be wrong (and get
  low score on test or project!)

23
Qualitative Value Model (Value Hierarchy)

• There can be multiple tiers or levels so that
  the problem can be subdivided into small enough
  pieces
• At the bottom tier are the attributes or
  measures or metrics that are used to quantify
  achievement in the objectives
• Use terms most appropriate for the domain

24
Objective Hierarchy Example

• Building / Buying a computer
• Convert that hierarchy to an objective hierarchy
• Decision Context (value)
• Measure
• Direction of preference

25
Measures of Effectiveness (MOEs)

• MOEs are
• Related to the objective
• Measurable and quantifiable
• (by convention) bigger is better
• In military analysis, we often distinguish
  between
• Measures of Performance (MOPs)
• Measures of Effectiveness (MOEs)
• Measures of Force Effectiveness (MOFEs)
• Sometimes the general term Measures of Merit
  (MOMs) are used to generically describe MOPs,
  MOEs, and MOFEs

26
Dealing with Multiple MOEs

• Frequently we deal with multiple objectives
• E.g., a ship must be
• Survivable
• Fast
• Carry certain weapons platforms
• Self-sustaining for a certain period of time
• Associated with multiple objectives are multiple
  MOEs.
• E.g.
• Flank speed of X knots
• Ability to float with Y number of compartments
  holed
• Carry Z1 TLAMs and Z2 SM-2s
• Decisionmakers like to compare things on a single
  scale
• Is A better than B?

27
Dealing with Multiple Objectives / MOEs

• The same as comparing apples and oranges
• Different attributes and different opinions about
  values
• Sometimes you dont want fruit salad!
• One approach Convert everything to a common
  single scale
• Examples
• Dollars (money)
• Easy to do when comparing business alternatives
  when are a natural measure
• Hard to equate some things with money - what is
  the equivalent of a human life? (yet it is done
  every day)
• Utility
• Rules and Axioms hold (we will cover them)
• Utility is inherently subjective - Value is in
  the eyes of the beholder
• The hard part is generally eliciting utility
  curves from people

28
When

• Suppose you were offered a choice
• 2 today and 2 tomorrow, or
• 4 today
• They are the same, right? You have 2 2 4
• The answer, of course, is NO !

The value of money depends on when you have it.
2 today gt 2 tomorrow Why do we care?
29
The Time Value of Money

• Assumption If we have x dollars today, we can
  earn a return of r percent on investment over 1
  time period

Suppose the time period is a year, and r 3.
Then after 1 year, the x dollars you have today
become x r dollars tomorrow. After n years, we
get dollars.
Alternatively, we could say that the Present
Value (PV) of x dollars n years from now is
30
Creating Alternatives

• No recipe but Values First!
• First alternative thought of
• Readily available
• Used before
• Constrained
• Includes the Status Quo
• Find alternatives that meet objectives
• Consider objectives in groups (2 or 3 at a
  time)
• Focus on strategic objectives
• Most important
• Most broad
• Will do most good

31
We usually need better alternatives.
Values What do we want?
Future What could happen?
Information What do we know?
Alternatives What can we do?
32
Identifying better alternatives is a key role for
decision analysts.
If you want better decisions,
find better alternatives!
If there are nothing but bad alternatives and you
just evaluate alternatives, all your help and
analysis focuses on determining the best bad
alternative!

33

There are barriers to finding better alternatives.

• Using Alternative-Focused
• Thinking (Status Quo)
• Environment that
• discourages creativity
• Not finding smart, creative people
• Lack of good alternative generation techniques
• Too many alternatives to evaluate

Better Alternatives
34
Questions on values?