How to Make A Decision:

1
Really Good

• How to Make A Decision
• Applications to Engineering Trade Studies
• Dr. Steve Walter
• Department of Engineering
• October 23, 2006

2
Table of Contents

• Attributes of a good decision
• Decision making process
• Requirements and goals
• Importance of stakeholders
• Defining alternatives
• Ranking goals
• Figure of merits
• Quantifying risk
• Comparison of alternatives
• Making the choice
• Documenting it.
• Summary

3
Attributes of a Good Decision

• Rely on well-defined stakeholder preferences
• Based on objective information
• Consider all alternatives
• Quantitative
• Balanced against risk
• Transparent
• Documented

4
The Route to a Good Decision
Define Decision
Establish Requirements
Generate Alternatives
Rank Goals
Identify Goals
Document Decision
Make Choice
Identify and Assess Risk
Compare and Contrast
5
Requirements and Goals

• Requirements are needs or musts
• Good requirements are
• Clear
• Concise
• Attainable
• Verifiable
• Terminology Requirements use shall
• Usually defined with inequalities
• Goals are desirements or wants
• Good goals are
• Clear
• Concise
• Not impossible
• Verifiable
• Terminology Goals use should.
• Usually defined with inequalities

Requirements determines who races, Goals
determines who come in first.
6
An Example Choosing a Family Sedan

• Good Requirements and Goals for a Family Car
• Pick and quantify the requirements
• It shall be a sedan
• The price shall be less than 30,000
• It shall have four doors
• It shall have cruise control
• Pick and quantify the goals
• The in-city gas mileage should be 25 mph
• It should have a in-dash navigation system
• It should have leather seats
• It should have frontal and side impact NHTSA five
  star rating
• It should have a trunk that is 14 cu. ft.
• It should have five-axis seat adjustments

• Needs and Wants for a Family Car
• The car must be a sedan
• The car must not be too expensive
• It must have four doors
• It must get good gas mileage
• It must have cruise control
• It must have a in-dash navigation system
• It must have leather seats
• It must be safe
• It must have a large trunk
• It must be comfortable

7
Information Gathering is Key
The car must be comfortable
What?

• It should have five-axis seat adjustments
• It should have at least 36 inches of head room
• It should have at least 39 inches of leg room
• It should be able to seat five adults
• It should have MacPherson struts

8
Stakeholders Determine Requirements and Goals
Manufacturing
Users
Developers
9
Engineering Requirements and Goals

• Cost
• Development Costs (NRE)
• Manufacturing Cost (RE)
• Product Cost
• Product Price
• Lifecycle Costs
• Cost Risk

• Time
• Development Schedule
• Production / Mfg Time
• Schedule Risk

10
Develop Alternatives Choosing a Space
Antenna

• Three types of antennas meet the requirements for
  a future space mission

Deployable Cassegrain
Fixed Cassegrain
Off-axis Parabola

• Large Aperture
• Lightweight
• Moderate Sidelobes

• Low Cost
• No Development Risk
• No Deployment Risk

• Excellent Sidelobes
• Moderate Cost
• No Deployment Risk

11
Ranking Goals

• Assessing goal importance and assigning weights
  are the next step
• The Kepner and Tregoe ranking system
• 1-10 scale
• Pick the most important goal and rank it a 10.
• Then compare the next most important set of goals
  and see if they deserve a 10, if not give them
  a 9 or less.
• Work through all goals until each of them have a
  ranking
• Pie chart
• Create a pie chart and assign a goal to each
  wedge
• Adjust the size of each wedge until the area of
  each wedge is consistent with the goals
  importance
• The Size of each wedge is the weighting
• Key stakeholders use a consensus process to
  establish goal weights

12
Goal Rankings

• Antenna Goals
• The effective aperture should be greater than 90
  sq ft.
• The sidelobes should be less than -30 db
• The deployment risk should be low
• The cost of development should be less than 15M
• The antenna should weigh less than 40 lbs
• The list of goals are distributed to the
  customer and system engineering team to
  determine their ranking

13
Quantifying How Well an Alternative Meets a Goal

• Figure of merit translates measurable attributes
  into a number
• Ranges from 1 to 10
• Requires clear and verifiable goals
• Spans from the requirement to (or through) the
  goal
• Yes/no choices Yes 1 and No 0

14
Risk 101

• Risk
• The cumulative effect of uncertain occurrences
• that may negatively or positively affect
  objectives
• Uncertainty
• An indefinite or indeterminate quantity
• Reports that say that something hasn't happened
  are always interesting to me, because as we know,
  there are known knowns there are things we know
  we know. We also know there are known unknowns
  that is to say we know there are some things we
  do not know. But there are also unknown unknowns
  -- the ones we don't know we don't know.
• Donald Rumsfeld

Risk management deals with known unknown
15
Risk Examples 102

• Examples of Risk Events
• Technical
• Piston rings do not hold specified pressure
• Power supply has wrong connector
• Deployable antenna doesnt open
• Organizational
• Key people are not available to work on project
• Planned test or mfg equipment is not available
• External
• Customer cancels contract due to convenience
• Lightning strike (i.e., Verizon) or Earthquake
• Government safety standards change
• Printer breaks when you need to print the final
  report
• Purchases
• Vendor does not deliver a product on time
• Subcontractor key staff quits or goes out on
  strike
• (Government assigns 20 - 40 weight to past
  performance.)

Galileos antenna did not deploy limiting data
return
16
Quantifying Risk 103

• Risk has three components
• Description of the risk event
• Probability of occurrence or likelihood
• Severity or impact
• Can be scored on a five-point scale
• Quantified in dollars, days of delay, units of
  performance, etc.

Likelihood 1 5 5 40 40
Cost Severity 1 minimal impact 2 5
overrun 5 10
overrun
Power Severity 1 No impact 2 2 HP 2 HP
6 HP
17
Risk 104

• Computer Risks
• Disk drive crash
• Likelihood 2, Severity 5
• Product 10, Normalized 6
• Existing software cant read the file format
• Likelihood 3, Severity 2
• Product 6, Normalized 7.6
• Network connection goes down
• Likelihood 4, Severity 3
• Product 12 Normalized 5.2
• Computer crash causes loss of active files
• Likelihood 2, Severity 2
• Product 4, Normalized 8.4
• Flight attendant spills a drink into your laptop
  frying the keyboard
• Likelihood 1, Severity 5
• Product 5, Normalized 8

Likelihood
1 2 3 4 5
1 2 3 4 5
Severity
18
Caveats Risk 105

• Risk Responses Lessen Severity and Likelihood
• Mitigate
• Identify a contingency that will reduce severity
  or likelihood
• Transfer
• Make the severity or likelihood someone elses
  problem
• Avoid
• Recast solution space to eliminate the problem
• Accept
• Risk Tolerance is Determined by Stakeholders

19
Determining the Figure of Merit Choosing a
Space Antenna




Total Figure of Merit
20
Space Antenna Options Comparison
Total Figures of Merit
21
Adverse Consequences

• If risk has been explicitly included in the
  determining the figure of merit it needs to be
  incorporated after the comparison
• Removing Risk from the equation we get

• Risk Tolerance is determined by stakeholders

After considering adverse consequences a choice
can be made
22
Documentation

• Documented items
• Date of decision
• Decision maker
• Problem or decision statement
• Requirements and goals
• Considered alternatives
• Goal weights and rankings
• Comparison
• Risks
• Choice

A documented problem/decision statement
complete with a description of alternatives
The decision that has not been documented is not
a decision
23
Summary Attributes of a Good Decision

• Rely on well-defined stakeholder preferences
• Based on objective information
• Consider all alternatives
• Quantitative
• Balanced against risk
• Transparent
• Documented

24
One university.Two great names.www.ipfw.edu