BE 20 Engineering Design with Computer Applications

1
Concept Selection Mathematical Models for
Decision-Making

• BE 20 - Engineering Design with Computer
  Applications
• Week 8 13-October-2004

2
Todays Journey

• Due today Design Report 1
• Announcements
• Stages in Team Development
• Where is your team today?
• Using a Decision Matrix to Evaluate Solutions
• Estimation of parameters
• Decision Matrix
• Reminders and Assignments

3
Announcements

• BE20 Fab Lab opens 18-Oct-2004
• Hours (starting 20-Oct)
• M-F 8 AM - 4 PM
• M-Th 4-10 PM (Lab monitor on duty)
• Shop Safety on 15, 18-Oct-2004
• Come prepared to use tools
• No dangling jewelry or open toe shoes

4
Team Development Stages
ADJOURNING
PERFORMING
NORMING
STORMING
FORMING
5
Team Development Stages FORMING STAGE
ADJOURNING
PERFORMING

• Common behaviors
• politeness
• uncertainty
• attempting to define tasks
• determining acceptable behavior
• deciding where to begin
• diving into solutions
• depending on single individual

NORMING
STORMING
FORMING
6
Team Development Stages STORMING STAGE
ADJOURNING
PERFORMING

• Common behaviors
• arguing
• defensiveness
• setting unrealistic goals
• choosing sides
• passing blame
• being rude

NORMING
STORMING
FORMING
7
Team Development Stages NORMING STAGE

• Common behaviors
• striving for harmony
• openly expressing opinions
• sharing information
• less resistance to team tasks

8
Team Development Stages PERFORMING STAGE
ADJOURNING

• Common behaviors
• balancing contributions
• focusing on goals
• solving problems together
• reaching consensus
• sharing accountability
• following through
• pushing for higher standards

PERFORMING
NORMING
STORMING
FORMING
9
Team Development Stages ADJOURNING STAGE
ADJOURNING

• Team has accomplished task
• Common behaviors
• prepare to disband
• members feel regret/loss
• grieve
• act out

PERFORMING
NORMING
STORMING
In what developmental stage is your team now?
FORMING
10
Case Study Portable Bridge

• The setup
• A need was identified for a small, portable
  bridge that could span a 7 ft wide chasm. The
  bridge needs to be lightweight (so it can be
  carried to a remote location), small (in bulk),
  quick to set up and assemblable from one side of
  the chasm.
• Three candidate designs have been proposed
• We now must choose only one to build

11
Case Study (2)

• Bridge concept Alpha Telescoping bridge
• A compact concept where sections are nested
  inside each other. A spring loaded button holds
  the sections in place when expanded. Material is
  0.125 Al. Longer than the required 7 ft.

12
Case Study (3)

• Bridge concept Beta Ladder bridge
• Lightweight concept that requires little assembly
  at the river site. Uses one-way hinges to
  support load. Material is lightweight metal.
  Metal slats provide walkway.

13
Case Study (4)

• Bridge concept Gamma Plastic frame
• A sectional bridge concept made of PVC piping.
  Sections are connected with PVC couplings and
  cross-sections. Cross-sections provide support
  for walkway of plastic milk crate mesh.

14
Case Study (5)

• How do you decide which one to build?

15
Design Process Overview

• To clarify the design problem, we created
• Objectives tree
• Functional model
• To generate concepts, we created
• Morphological chart
• Concept variants

16
Design Process Overview (2)

• Where we are going
• Select a single concept to fabricate (today)
• Our approach Concept selection by Decision
  Matrix
• Build proof of concepts to test critical
  components of the overall solution
• Fabricate a prototype of the chemical powered
  vehicle for testing/competition

17
Concept Selection Techniques

• Basic concept selection
• Comparison of concept variants to identify
  feasible and infeasible ones
• The next step Compare concept variants to
  determine the best one to fabricate
• Based on qualitative and quantitative reasoning
• FOM calculations are quantitative
• Other criteria, such as product appearance, are
  more qualitative in nature
• For the bridge case study example
• Quantitative objectives Deflection of the bridge
  at center span, weight, time to set up
• Qualitative objectives Appearance,
  innovativeness

18
Why use a Decision Matrix? (a.k.a. Concept
Selection Matrix)

• From the Morphological Chart, you
• Generated a large number of solutions
• Assembled concept variants or CVs
• How do we evaluate these solutions? What makes
  one better than another?
• Knowing how to choose between alternatives is an
  important design activity!
• You can choose by some arbitrary manner, but it
  is best to use a rational approach that can be
  clearly explained to managers and clients
• What makes a Decision Matrix a rational approach?
• Weighted numerical scores are assigned to various
  sub-objectives
• Individual scores are summed to give each overall
  solution choice a numerical score

19
How to Create a Decision Matrix
General Procedure
Step

• 1. List the design objectives
• Use your Objectives Tree
• You may need to modify the initial list somewhat
• 2. Rank order these from most to least important
• For example, when ordering a pizza, getting
  pepperoni is more important than how much cheese
  or the crust type
• A rank order is a list from most important to
  least, but it may appear to be a linear scale,
  which it often is not some objectives are much
  more important than others

20
How to Create a Decision Matrix (2)
General Procedure
Step

• 3. Assign relative weighting to these objectives
• Some objectives are much more important than
  others
• For example, having pepperoni is a 10, enough
  cheese is a 6, crust type is a 1
• This is best done as a team
• 4. Rate each concept variant against each design
  objective
• 5. Sum up the weighted ratings for each concept
  variant and compare
• In general, higher values mean better concept
  variant

21
FOM-based Decision Matrix
Project Application Steps 1-3

• The Figure of Merit (FOM) for your chemical
  powered vehicle
• FOM 10(50-C) 10(240-D) 1.3V 5(180-R)
  50(120-T) 20?A 40?DI 30?OI 30?EI
• C Cost of materials ()
• D Measured distance from front of vehicle to
  finish line (in.)
• V Volume of water carried by vehicle, up to
  1000 mL (mL)
• R Reset time (s)
• T Time to complete course (s)
• A Assessment of workmanship and aesthetic
  appeal of the device, including ease and
  reliability of operation (0 A 4)
• DI Assessment of the creativity and/or
  innovation of the design (0 DI 4)
• OI Assessment of the quality of the operating
  instructions (0 OI 4)
• EI Assessment of the environmental impact of
  the device (0 EI 4)

Decision matrix objectives
22
Weighting Factors
Project Application Steps 1-3

• The given Figure of Merit (FOM) for your chemical
  powered vehicle also includes weighting factors

23
Estimating parameters
Project Application Step 4

• To compare potential chemical powered vehicle
  designs, your team will need to predict the
  Figure of Merit (FOM) for each concept variant
• Do this for all five (or more) concept variants
  generated from your Morphological Chart
• For each concept variant, estimate the parameters
  that appear in the FOM (cost, distance, volume,
  reset time, time, )
• Base these estimates on common sense, your
  existing knowledge, calculations, research on
  web, catalogs, etc.
• REMEMBER If your estimates are bad, then the
  resulting decision matrix results will be garbage
• Enter the estimates in a table

24
Estimating Parameters for Your Project
Project Application Step 4
Enter the raw (i.e., unweighted) value of the
objective for each concept variant
Continue estimating parameters as accurately as
you can until the entire table is completed
25
FOM Decision Matrix
Project Application Step 5

• Take the parameters you estimated for each
  solution and create an Excel spreadsheet like
  this to determine a score for each of your
  concept variants
• Do this for ALL of your concept variants
• The total row gives you a numerical score for
  each of your potential solutions
• Which solution is best?

26
FOM Decision Matrix
Project Application Step 5

• For example, the overall decision matrix (in
  excel) might look like

27
Interpreting the Result

• Back to the bridge case
• Which one do we build?

28
Interpreting the Result (2)

• Main point NEVER blindly accept the results of a
  decision matrix
• Use common sense!
• If two CVs have relatively close scores, treat
  them as being tied
• You will need to further discriminate between the
  two!
• If results are unexpected, ask whether
  expectations were wrong, whether the method was
  applied consistently or whether the weights are
  not appropriate
• If results meet expectations, ask whether they
  represent a fair application of the process or if
  you biased the result to select a preconceived
  concept variant
• Finally, if some alternatives were rejected due
  to constraint violations, double check if those
  constraints are truly binding

29
Interpreting the Result (3)

• Back to the bridge case
• What would you do now?
• How could you further discriminate between CV1
  and CV3?

30
Further Discrimination

• Our FOM-based decision matrix is a simple
  starting point for concept selection
• Other methods include
• Pugh charts - these use one CV as a baseline and
  compare all others to it
• Decision matrices incorporating other objectives
  (other than simply the FOM terms)
• If the FOM consistently gives you two or more CVs
  with the same score, consider
• Adding in other objectives to discriminate
  between close scoring CVs (such as manufacturing
  difficulty/cost)
• Reverifying your parameter estimates

31
Reminders and Assignments

• Memo 6 (Due Wed, 20-Oct-2004)
• Create a FOM-based decision matrix for your
  chemical powered vehicle as shown in this
  presentation
• Use Excel to calculate decision matrix values
• Include sketches of your concept variants for
  reference
• Attach supporting analyses for parameter
  estimations
• Attach above items to a cover memo and describe
  the best solution, outlining the solution
  approach you will use
• Next time Developing Models and Prototypes

32
Questions?