1 ? Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Design for Value
Lecture 24
10 May 2004
Karen Willcox
Olivier de Weck
Acknowledgments: Jacob Markish and Ryan Peoples
Multidisciplinary System
Design Optimization (MSDO)
2 ? Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Today’s Topics
An MDO value framework
Lifecycle cost models
Value metrics & valuation techniques
Value-based MDO
Aircraft example
Spacecraft Example
3 ? Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Optimal Design
Traditionally, design has focused on performance
e.g. for aircraft design
optimal = minimum weight
Increasingly, cost becomes important
85% of total lifecycle cost is locked in by the end of
preliminary design.
But minimum weight z minimum cost z maximum value
What is an appropriate value metric?
4 ? Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Design Example
We need to design a particular portion of the wing
Traditional approach: balance the aero & structural requirements,
minimize weight
We should consider cost: what about an option that is very cheap to
manufacture but performance is worse?
aerodynamics?
How do we trade performance and cost?
How much performance are we willing to give up for $100 saved?
What is the impact of the low-cost design on price and demand of
this aircraft?
What is the impact of this design decision on the other aircraft I
build?
What about market uncertainty?
structural dynamics?
manufacturing cost?
aircraft demand?
aircraft price?
tooling?
environmental impact?
5 ? Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Cost
Module
“Value” metric
Performance
Module
Aerodynamics
Structures
Weights
Mission
Stability & Control
Revenue
Module
Value Optimization Framework
Manufacturing
Tooling
Design
Operation
Market factors
Fleet parameters
Competition
“Optimal”
design
6 ? Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Challenges
Cost and revenue are difficult to model
– often models are based on empirical data
– how to predict for new designs
Uncertainty of market
Long program length
Time value of money
Valuing flexibility
Performance/financial groups even more uncoupled
than engineering disciplines
7 ? Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Cost Model
Need to model the lifecycle
cost of the system.
Life cycle :
Design - Manufacture -
Operation - Disposal
Lifecycle cost :
Total cost of program over
life cycle
85% of Total LCC is locked
in by the end of preliminary
design.
Cost
Module
“Value” metric
Performance
Module
Revenue
Module
8 ? Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Lifecycle Cost
0
20
40
60
80
100
65%
Conceptual
design
Preliminary design, system integration
Detailed design
Manufacturing and acquisition
Operation and support
Disposal
Time
Impact on LCC (%)
85%
95%
(From Roskam, Figure 2.3)
9 ? Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Non-Recurring Cost
Cost incurred one time only:
Engineering
- airframe design/analysis
- configuration control
- systems engineering
Tooling
- design of tools and fixtures
- fabrication of tools and fixtures
Other
- development support
- flight testing
Engineering
Tooling
Other
10 ? Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Development Cost Model
Cashflow profiles based on beta curve:
Typical development time ~6 years
Learning effects captured – span, cost
11
)1()(
E D
tKttc
0
0.01
0.02
0.04
0.05
0.06
1 2 3 4 5 6 7 8 9 1011121314151617181920212223242526272829303132333435363738394041424344454647484950515253
normalized time
Support
Tool Fab
Tool Design
ME
Engineering
normalized cost
(from Markish)
11 ? Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Recurring Cost
Cost incurred per unit:
Labor
- fabrication
- assembly
- integration
Material to manufacture
-raw material
- purchased outside
production
- purchased equipment
Production support
-QA
- production tooling support
- engineering support
Labor
Material
Support
12 ? Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Learning Curve
As more units are made, the recurring cost per
unit decreases.
This is the learning curve effect.
e.g. Fabrication is done more quickly, less
material is wasted.
n
x
xYY
0
Y
x
= number of hours to produce unit x
n = log b/log 2
b = learning curve factor (~80-100%)
13 ? Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Learning Curve
0.55
0
0.2
0.4
0.6
0.8
1
1020304050
Unit number
C
o
s
t
of
uni
t
b=0.9
Typical LC slopes: Fab 90%, Assembly 75%, Material 98%
Every time
production
doubles,
cost is
reduced by
a factor of
0.9
14 ? Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
CASH AIRPLANE RELATED
OPERATING COSTS:
Crew
Fuel
Maintenance
Landing
Ground Handling
GPE Depreciation
GPE Maintenance
Control & Communications
Airplane Related Operating Costs
CAROC is only 60% - ownership costs are significant!
CAROC
60%
40%
Capital
Costs
CAPITAL COSTS:
Financing
Insurance
Depreciation
15 ? Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Value Metric
Need to provide a
quantitative metric that
incorporates cost,
performance and
revenue information.
In optimization, need to
be especially carefully
about what metric we
choose...
Cost
Module
“Value” metric
Performance
Module
Revenue
Module
16 ? Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
What is Value?
Objective function could be different for each stakeholder
e.g. manufacturer vs. airline vs. flying public
Program related parameters vs. technical parameters
cost, price, production quantity, timing
Traditionally program-related design uncoupled from
technical design
Customer
Value
Shareholder
Value
Product
Quality
Schedule
Cost
Economic
Value
Added
Demand
Revenue
EBIT
System
Design
Price
From Markish,
Fig. 1, pg 20
Customer value derived from
quality, timeliness, price.
Shareholder value derived
from cost and revenue, which
is directly related to customer
satisfaction.
17 ? Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Value Metrics
performance
weight
speed
Traditional Metrics
cost
revenue
profit
quietness
emissions
commonality
...
Augmented Metrics
The definition of value will vary depending on your system
and your role as a stakeholder, but we must define a
quantifiable metric.
18 ? Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Valuation Techniques
Investor questions:
How much will I need to invest?
How much will I get back?
When will I get my money back?
How much is this going to cost me?
How are you handling risk & uncertainty?
Investment Criteria
Net present value
Payback
Discounted payback
Internal rate of return
Return on investment
19 ? Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Net Present Value (NPV)
Measure of present value of various cash flows in different
periods in the future
Cash flow in any given period discounted by the value of a
dollar today at that point in the future
– “Time is money”
– A dollar tomorrow is worth less today since if properly
invested, a dollar today would be worth more tomorrow
Rate at which future cash flows are discounted is
determined by the “discount rate” or “hurdle rate”
– Discount rate is equal to the amount of interest the
investor could earn in a single time period (usually a
year) if s/he were to invest in a “safer” investment
20 ? Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Discounted Cash Flow (DCF)
Forecast the cash flows, C
0
, C
1
, ..., C
T
of the project
over its economic life
– Treat investments as negative cash flow
Determine the appropriate opportunity cost of capital
(i.e. determine the discount rate r)
Use opportunity cost of capital to discount the future
cash flow of the project
Sum the discounted cash flows to get the net present
value (NPV)
NPV C
0
C
1
1 r
C
2
1 r
2
!
C
T
1 r
T
21 ? Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
DCF example
Period Discount Factor Cash Flow Present Value
0 1 -150,000 -150,000
1 0.935 -100,000 -93,500
2 0.873 +300000 +261,000
Discount rate = 7% NPV = $18,400
22 ? Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Risk-Adjusted Discount Rate
DCF analysis assumes a fixed schedule of cash flows
What about uncertainty?
Common approach: use a risk-adjusted discount rate
The discount rate is often used to reflect the risk
associated with a project: the riskier the project, use a
higher discount rate
Typical discount rates for commercial aircraft programs:
12-20%
Issues with this approach?
23 ? Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Net Present Value (NPV)
0
(1 )
T
t
t
t
C
NPV
r
|
-1500
-1000
-500
0
500
1000
1500
1 3 5 7 9
11 13 15 17 19 21 23 25 27 29
Cashflow
DCF (r=12%)
Program Time, t [yrs]
Cashflow,
P
t
[$
]
24 ? Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Payback Period
How long it takes before entire initial investment is
recovered through revenue
Insensitive to time value of money, i.e. no
discounting
Gives equal weight to cash flows before cut-off date
& no weight to cash flows after cut-off date
Cannot distinguish between projects with different
NPV
Difficult to decide on appropriate cut-off date
25 ? Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Discounted payback
Payback criterion modified to account for the time
value of money
– Cash flows before cut-off date are discounted
Overcomes objection that equal weight is given to
all flows before cut-off date
Cash flows after cut-off date still not given any
weight
26 ? Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Internal rate of return (IRR)
Investment criterion is “rate of return must be greater
than the opportunity cost of capital”
Internal rate of return is equal to the discount rate for
which the NPV is equal to zero
IRR solution is not unique
– Multiple rates of return for same project
IRR doesn’t always correlate with NPV
– NPV does not always decrease as discount rate
increases
NPV C
0
C
1
1 IRR
C
2
1 IRR
2
!
C
T
1 IRR
T
0
27 ? Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Return on Investment (ROI)
Return of an action divided by the cost
of that action
Need to decide whether to use actual or
discounted cashflows
revenue cost
cost
ROI
28 ? Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Decision Tree Analysis (DTA)
NPV analysis with different future scenarios
Weighted by probability of event occurring
29 ? Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Real Options Valuation Approach
?In reality:
?Cashflows are uncertain
?Ability to make decisions as future unfolds
?View an aircraft program as a series of
investment decisions
?Spending money on development today gives
the option to build and sell aircraft at a later
date
?Better valuation metric: expected NPV from
dynamic programming algorithm (Markish,
2002)
30 ? Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Dynamic Programming:
Problem Formulation
The firm:
– Portfolio of designs
– Sequential development phases
– Decision making
The market:
– Sale price is steady
– Quantity demanded is unpredictable
– Units built = units demanded
Problem objective:
– Which aircraft to design?
– Which aircraft to produce?
– When?
31 ? Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox
Engineering Systems Division and Dept. of Aeronautics and Astronautics
Dynamic Programming: Problem
Elements
1. State variables s
t
2. Control variables u
t
3. Randomness
4. Profit function
5. Dynamics
Solution:
Solve iteratively.
>