This document summarizes a presentation on using game theory, Bayesian belief networks, and quality function deployment (QFD) to support requirement development. It discusses business challenges like complex product development and lack of structure in allocating resources. Models were aligned with goals like simplifying initiatives and predicting impacts. Outcomes and factors were predicted and mapped for stakeholders using tools like Excel and HUGIN. Challenges included subjectivity and unclear roles, and what worked well was collecting historical data to build a generic model.
1. Game Theory
and
Bayesian Belief Network
to Support
Quality Function Deployment (QFD)
for Requirement Development
Kobi Vider
K.V.P Consulting
Kobi.vider@hotmail.com
2. Agenda
• Introduction
• Methods Overview
• Business Challenges
• Business Goals
• Goal Alignment with Models
• Outcome(s) Predicted
• Stakeholder Audience
• Factors used in the Process Performance Model
• Tool Used
• Challenges
• What Worked Well
• Summary
• Additional Posts
3. Introduction
• Two and a half years ago I Was asked to develop a method that will
support the initiation of complicated projects with large number of
overlapping stakeholders that influencing the system product
program scope, time and end deliverables.
• The baseline for evaluating what methods will be appropriate we did
postmortem and retro- perspective on five programs that ends
• The methods evaluation was conducted at different perspectives
(vertical and horizontal) including the use of the following tools:
Game Theory; Quality Function Deployment; Bayesian Networks
and Dynamic Bayesian Games
• This presentation is a brief summery of the process elements that we
were able to identify and the building parameters for its performance
measurements
• We will include in the presentation (as time will permit it) tools walk
through; I am willing to share it and send it upon request
4. Agenda
• Introduction
• Methods Overview
• Business Challenges
• Business Goals
• Goal Alignment with Models
• Outcome(s) Predicted
• Stakeholder Audience
• Factors used in the Process Performance Model
• Tool Used
• Challenges
• What Worked Well
• Summary
• Additional Posts
6. Introducing Example
• You are going to have a blind date
• but you don’t know where you will meet the other person
• Only if you two choose the same location as a meeting
point the date will actually take place
• To make it a little easier…
assume there are only two places to go: „Pizza hut“ in Tel
Aviv and „Café place“ in the Jerusalem
• Where would YOU go?
7. Expected Utility Theory
• Developed by von Neumann & Morgenstern
(1947)
• In short: The option with the greatest utility is
chosen
• Based on the three assumptions (axioms):
• Completeness: If there are 2 alternatives, an agent will prefer A or B
or is indifferent between A and B
• Transitivity: If an agent prefers A over B and B over C, he will also
prefer A over C
• Context-free ordering: If an agent prefers A over B, he will still
do this when additional alternatives (C, D, …) are available
8. Game Theory
• Is an idealized abstraction of reality
• Is a normative, not a descriptive theory
• It states only how people should behave if they wish to maximize
their utility
• It does not describe how people actually behave
• Can be tested empirically
• Experimental gaming experiments
9. Game Theory: Assumptions
• There are two assumptions of Common
Knowledge and Rationality (CKR)
• CKR 1: The specification of the game (e.g.
number of players, payoff functions) are known to
all players
• CKR 2: All players are rational in the sense of
Expected Utility Theory
All players will choose strategies that will maximize
their individual expected utilities
10. Team Reasoning
• Explanation for cooperative behavior in social
dilemmas
• A team reasoning player…
• maximizes the collective payoff
• chooses not by individual but by collective preference
• violates the second assumption of Common Knowledge
and Rationality on which game theory is based upon
• CKR 2: “All players will choose strategies that will maximize
their individual expected utilities”
11. Summary and Conclusion
• Classical Game Theory…
• is a normative theory based on Expected Utility Theory
• is not able to predict decisions in all interactive situations but
sometimes remains indetermined
and…
• predicts self-defeating behavior in social dilemmas
• Conclusion: Classical Game Theory is useful to understand
social interactions but needs to be modified
13. Background
• There has been little emphasis on the decision
making process on project program scoping.
• Scoping Rationale (SR) as a result of a decision is
often not clearly captured
• Even when SR is captured, it is often difficult to
explain how decisions relate to and affect the
architecture considerations
• Change impact cannot be systematically reasoned
or explained during the different phases
• It is difficult to quantify the impact of changes in
requirement, design or decision
14. Project Product Scoping Decisions
• Early Decision Making
• Consider Multiple Factors including Functional
Requirements, Non-Functional Requirements
and Environmental Factors
• Consider Different Perspectives and Viewpoints
• Directly and Indirectly Influence the Design
Structure of the System
• Create / Modify Design Elements to Satisfy
System Goals / Sub-goals
15. What Is Decision Analysis?
• Decision Analysis Provides Effective Methods for
Organizing a Complex Problem into a Structure that
can be Analyzed
• Identifies Possible Courses of Action
• Identifies Possible Outcomes
• Identifies Likelihood of the Outcomes
• Identifies Eventual Consequences
• Decision Analysis Provides the Methods to Trade
Off Benefits Against Costs
• Decision Analysis Allows People to Make Effective
Decisions More Consistently
16. Problem Statements
• How to capture rationale and represent
architecture considerations related decisions
in relation to design artefacts?
• What is the change impact to the system
when one or more requirements, designs or
decisions are to change?
• Tool walk through
17. Game Theory
and
Bayesian
Static Bayesian Games
Multi-stage games
Dynamic Bayesian Games
18. What is Bayesian Game?
Game in strategic form
- Complete information(each player has perfect information
regarding the element of the game)
- Iterated deletion of dominated strategy, Nash equilibrium:
solutions of the game in strategic form
Bayesian Game
- A game with incomplete information
- Each player has initial private information,
- Bayesian equilibrium: solution of the Bayesian game
19. Static Bayesian Games
• Static Games of Incomplete Information
• In many economically important situations the game may begin
with some player having private information about something
relevant to her decision making.
• These are called games of incomplete information, or Bayesian
games. (Incomplete information is not to be confused with
imperfect information in which players do not perfectly observe
the actions of other players.)
• Although any given player does not know the private information
of an opponent, he will have some beliefs about what the opponent
knows, and we will assume that these beliefs are common
knowledge.
• In many cases of interest we will be able to model the
informational asymmetry by specifying that each player knows her
own payoff function, but that she is uncertain about what her
opponents' payoff functions are
20. Important class of signaling games
• In signaling games there are two players, Sender
and Receiver
• Type of Sender is private information, sender
takes an action
• Strategy is action depending on type
• Receiver takes an action after observing action
taken by the sender
• Type of sender may be inferred (revealed) on the
basis of the action that is actually taken
22. Definition of Quality
Function Deployment :
There is no single, right definition for QFD; this one
captures its essential meaning:
A system for translating customer requirements into
appropriate company requirements at each stage from
research and product development to engineering and
manufacturing to marketing/sales and distribution
Prerequisites to QFD are ‘Market Research’ and ‘VOC gathering’.
As QFD is the process of building capability to meet or exceed customer demands,
understanding the market, knowing the various customer segments. what each customer
segment wants, how important these benefits are, and how well different providers of
products address these benefits are some of the key precursors to a successful QFD.
These are prerequisites because it is impossible to consistently provide products /
services which will attract customers unless you have a very good understanding of what
they want.
23. Kano Customer Need Model
Where does QFD fit?
Go to Tool
• UNEXPECTED,
PLEASANT SURPRISES
• 3M CALLS THEM
Satisfied CUSTOMER DELIGHTS
Customer
Spoken
Measurable
Range of Fulfillment
Excitement
Needs
QFD focuses on
Performance
Don’t Have Included
Don’t Do Do Well
Needs and unmet
Basic Needs
Unspoken
Performance Taken For granted
Basic
Needs Spoken If Not Met
Basic
Needs Dissatisfied
Customer
RECOGNIZE 1) The Impact of Needs on the Customer
2) That Customer Needs Change With Time
3) The impact of Communication of Customer Wants
Throughout the Organization
24. Quality Function Deployment’s
House of Quality Correlation 6
Matrix
3
Design
Attributes
Importance Rankings
2 5
1
Customer Relationships Customer
4
Needs between Perceptions
Customer Needs
and
Design Attributes
7
Costs/Feasibility
8
Engineering Measures
25. Agenda
• Introduction
• Methods Overview
• Business Challenges
• Business Goals
• Goal Alignment with Models
• Outcome(s) Predicted
• Stakeholder Audience
• Factors used in the Process Performance Model
• Tool Used
• Challenges
• What Worked Well
• Summary
• Additional Posts
26. Business Challenges
• Complex Product Development Initiatives
• Communications Flow Down Difficult
• Expectations Get Lost
• New Product Initiatives / Inventions
• Lack unclear Structure or Logic to the Allocation of Development
Resources.
• Large Complex or Global Teams
• Challenges in processes efficiency And/or Effectiveness
• Teamwork coordination Issues
• Conflicts in Product Development Times
• Excessive Redesign
• Changing Teams
• Problem Solving, or Fire Fighting.
27. Agenda
• Introduction
• Methods Overview
• Business Challenges
• Business Goals
• Goal Alignment with Models
• Outcome(s) Predicted
• Stakeholder Audience
• Factors used in the Process Performance Model
• Tool Used
• Challenges
• What Worked Well
• Summary
• Additional Posts
28. Business Goals
• Simplified the Product Development Initiatives to clear scope and users
• Identify, map and assign appropriate priorities the different stakeholders and
commitments
• Identify and predict the New Product Initiatives / Inventions impact on the
program and other stakeholders
• Identify and predict the Large Complex or Global Teams coordination and
alignment efforts Inventions impact on the program and other team members
teams
• Identify and predict processes efficiency And/or Effectiveness impact on the
program and teams
• Identify and predict Conflicts in Product Development Time vs. the
stakeholders expectations
• Identify and predict redesign Effectiveness impact on the program and teams
• Identify and predict changing in teams impact on the program and teams
• How to choose the right way Problem Solving, or Fire Fighting based on
quantitative and prediction of impact analysis
29. Agenda
• Introduction
• Methods Overview
• Business Challenges
• Business Goals
• Goal Alignment with Models
• Outcome(s) Predicted
• Stakeholder Audience
• Factors used in the Process Performance Model
• Tool Used
• Challenges
• What Worked Well
• Summary
• Additional Posts
30. Goal Alignment with Models - 1
• Simplified the Product Development Initiatives to clear
scope and users
• QFD and Dynamic Bayesian Games
• Identify, map and assign appropriate priorities the different
stakeholders and commitments
• Quality Function Deployment
• Identify and predict the New Product Initiatives /
Inventions impact on the program and other stakeholders
• Game Theory; Bayesian Networks and Dynamic Bayesian Games
• Identify and predict the Large Complex or Global Teams
coordination and alignment efforts Inventions impact on
the program and other team members teams
• Bayesian Networks and Dynamic Bayesian Games
31. Goal Alignment with Models - 2
• Identify and predict processes efficiency And/or Effectiveness impact
on the program and teams
• Bayesian Networks and Dynamic Bayesian Games
• Identify and predict Conflicts in Product Development Time vs. the
stakeholders expectations
• Game Theory; Quality Function Deployment; Bayesian Networks and
Dynamic Bayesian Games
• Identify and predict redesign Effectiveness impact on the program and
teams
• Quality Function Deployment; Dynamic Bayesian Games
• Identify and predict changing in teams impact on the program and
teams
• Dynamic Bayesian Games
• How to choose the right way Problem Solving, or Fire Fighting based
on quantitative and prediction of impact analysis
• Bayesian Networks and Dynamic Bayesian Games
32. Agenda
• Introduction
• Methods Overview
• Business Challenges
• Business Goals
• Goal Alignment with Models
• Outcome(s) Predicted
• Stakeholder Audience
• Factors used in the Process Performance Model
• Tool Used
• Challenges
• What Worked Well
• Summary
• Additional Posts
33. Outcome(s) Predicted
• We have developed players stakeholders
map we have include the description of the
expected outcome(s) and its influence on
the ‘project’ performance
• The map template and example will be
uploaded to the website
34. Agenda
• Introduction
• Methods Overview
• Business Challenges
• Business Goals
• Goal Alignment with Models
• Outcome(s) Predicted
• Stakeholder Audience
• Factors used in the Process Performance Model
• Tool Used
• Challenges
• What Worked Well
• Summary
• Additional Posts
35. Stakeholder Audience
• We have developed a players stakeholders
map we have include the description of the
expected outcome(s) and its influence on
the ‘project’ performance, used to
communication and negotiations on
decisions
• The map template and example will be
uploaded to the website
36. Agenda
• Introduction
• Methods Overview
• Business Challenges
• Business Goals
• Goal Alignment with Models
• Outcome(s) Predicted
• Stakeholder Audience
• Factors used in the Process Performance Model
• Tool Used
• Challenges
• What Worked Well
• Summary
• Additional Posts
37. Factors used in the Process
Performance Model
• We include the factors in our data map tool
(e.g. influence )
• The map template and example will be
uploaded to the website
38. Data Collection
• Due to the unique nature of data elements and related factors we have
collected and analyzed the data elements and factors manually based
on players stakeholders per project program
• We have initiated historical data base (Excel based) and we are in the
progress to build generic model
• We did not use any sampling because for each project program we
need to run the full method from start, therefore we have developed
supporting matrix when to apply it
• The current threats to data quality and integrity that we have faced
• Players subjectivity
• Unclear player role
• Change of players (individuals) in the same position during one or more of
the ‘game’ (project program) instance
• We are currently running postmortem on past project to clean and
understand our percentage of measurement error
39. Agenda
• Introduction
• Methods Overview
• Business Challenges
• Business Goals
• Goal Alignment with Models
• Outcome(s) Predicted
• Stakeholder Audience
• Factors used in the Process Performance Model
• Tool Used
• Challenges
• What Worked Well
• Summary
• Additional Posts
40. Tool Used
Game Theory (Using Excel)
Bayesian Belief Network (Using HUGIN)
Quality Function Deployment (QFD) for
Requirement Development (Using Excel)
41. Agenda
• Introduction
• Methods Overview
• Business Challenges
• Business Goals
• Goal Alignment with Models
• Outcome(s) Predicted
• Stakeholder Audience
• Factors used in the Process Performance Model
• Tool Used
• Challenges
• What Worked Well
• Summary
• Additional Posts
42. Challenges
• How was the voice of the customer determined?
• How were the design requirements (etc) determined?
Challenge the usual in-house standards.
• How do we compare to our competition?
• What opportunities can we identify to gain a competitive edge?
• What further information do we need? How can we get it?
• How can we proceed with what we have?
43. Agenda
• Introduction
• Methods Overview
• Business Challenges
• Business Goals
• Goal Alignment with Models
• Outcome(s) Predicted
• Stakeholder Audience
• Factors used in the Process Performance Model
• Tool Used
• Challenges
• What Worked Well
• Summary
• Additional Posts
44. What Worked Well
• What worked well
• Senior staff commitments
• Stakeholders acceptance of the balancing results
• Stakeholders acceptance of their ‘position’ and weight
• Between our side benefits
• ‘snow ball’ effect from other departments
• Request for generic model development
• Request to adjust it to strategic and multi year programs
• Stakeholder inputs
• Give clear world view of all aspects
• Reduce the decision making and factors analysis complexity
• The historical data base from past projects reduce resistance
• Model development team member inputs
• Create more clear understanding on the
• The historical data base from past projects reduce development
time
45. Agenda
• Introduction
• Methods Overview
• Business Challenges
• Business Goals
• Goal Alignment with Models
• Outcome(s) Predicted
• Stakeholder Audience
• Factors used in the Process Performance Model
• Tool Used
• Challenges
• What Worked Well
• Summary
• Additional Posts
46. Summary - 1
• The process may look simple, but requires effort.
• Many of the entries look obvious - after they are written down.
• If there aren’t some “tough spots” the first time, it probably isn’t
being done right!
• Focus on the end-user customer.
• Charts are not the objective.
• Charts are the means of achieving the objective.
• Find reasons to succeed, not excuses for failure.
47. Summary - 2
• The combined methods that we have developed and use meant
to serve as the main support tool for the project program
Management, That Focuses on What the Customer Wants; and
Then Provides a Systematic Approach, Involving Communication
Between All stakeholders and Areas of the Organization, to
Make Sure These Wants Are Satisfied within the given
constraints.
• Decision Analysis Provides the Structure and Guidance for
Systematic Thinking
• Decision Analysis Process Organizes a Complex Problem into a
Structure that can be Analyzed
• Good Decision Analysis Requires Clear Thinking; Sloppy
Thinking Results in Worthless Analysis!
48. Agenda
• Introduction
• Methods Overview
• Business Challenges
• Business Goals
• Goal Alignment with Models
• Outcome(s) Predicted
• Stakeholder Audience
• Factors used in the Process Performance Model
• Tool Used
• Challenges
• What Worked Well
• Summary
• Additional Posts
49. Additional Posts
• Game Theory Overview Presentation
• Bayesian Belief Network Overview Presentation
• Quality Function Deployment (QFD) Overview
Presentation
• Players Stakeholders Map – Excel Based
• Data Type Map – Excel Based
• Bayesian Belief Network – HUGIN Based
http://www.hugin.com/Products_Services/Products/Demo/Lite/
• Quality Function Deployment - Excel Based