Lecture 1 Outline
• Organizational Information about the course
• Architectural Definitions and Role in ESD
• Systems Typollogy
ysis” Preview
• Network “Analy
• Pre-work for Lecture 2
Professor C. Magee, 2010
Page 1
Advanced System Architecture
ESD 342/EECS 6
ESD.342/EECS
6.883
883 2010
• Learning Objectives for this course:
• Gain
G i a researchh-llevell und
dersttandi
ding off systtem archit
hitectture
• Learn existing theoretical and analytical methods with
particular emphasis on network analysis
• Begin to develop some modeling skills of possible benefit
in complex engineering systems
• Compare
Co p e sys
systems
e s in ddifferent
e e domains
do
s (communication,
(co
u c o ,
engineering, organizations, infrastructures, and biology)
and understand what influences their architectures
• Apply/extend existing theory and modeling in case studies
Professor C. Magee, 2010
Page 3
Student Course Activities
• Read the academic literature, including faculty notes and
papers
p
• Learn and practice some existing analytical methods, mainly
network methods
• Appreciate the wide range of domains where theory and
methods have been applied
• Critique existing theory and methods
• Sh
Share knowlled
dge andd experiience
• Analyze some real systems in detail
• Distill common concep
pts that emergge from theoryy and that
apply to many kinds of systems
Professor C. Magee, 2010
Page 4
Learning Support for ESD 342
•
•
•
•
•
•
•
Class lecture/discussion
Text: Six Degrees by Duncan Watts (get a copy soon)
Additional Assigned Literature to read before class
Optional background literature available on the class web page
Two cllasses to learn
l
ab
bout thhe avail
ilable
bl sofftware.
Two homework assignments to learn to use the software
Case study
y ((small)) group
p pprojject with periodic repports in class
Professor C. Magee, 2010
Page 5
Grading Formula
• 15% in-class participation (especially reading
connections)
• 25% assignments
• 60% Project
• 20% Final Written Report
• 15% Final Presentation
• 15% Modeling Status Presentation
• 10% 1st Status Presentation
Professor C. Magee, 2010
Page 7
ESD’s Domains
"Systems Approach"
Architecture
Structure
and its
Relation to
Behavior
Networks
Hierarchies
Other Structural
Models
Decompositions
Complexity
Other Quantifications
of Architectures
Dynamics
Ilities
Architectural
Dynamic
Ethical, etc
Uncertainty
Ilities Related
to
Risk and
Uncertainty
Technical Ethics
"Extended" Economics
Enterprises
Policy
Human Behavior
Ilities Related to
Ethics
Sustainability
Safety
Extended Enterprises
Economic Ilities
S-curves, etc
Nations, firms
Bounded Rationality
Agency
Complexity and
Responses to it
Focus of this course
Professor C. Magee, 2010
Page 8
Why We Care
• Lotts off things
thi
have archit
hitecttures
• Physical things - objects, large natural systems
• Human designed things - products, systems, missions,
processes (engineering
( i
i andd others),
th ) organizations,
i ti
projects,
j t
infrastructures, software, databases, political and economic
systems
• Natural things - cells,
•
cells organisms,
organisms herds,
herds ecosystems
• Their architectures either determine, strongly influence, or are (at
least) correlated with their behavior, properties and changes over
time
• Understanding architectural progress and evolution is thus a key to
understanding systems over time (or at any given time)
• Thus,
Thus architecture has practical and fundamental importance for
engineering systems
Professor C. Magee, 2010
Page 9
A “Perfect” Theory of Architecture Would Permit Us To:
•
•
•
•
•
Characterize
Measure
Understand at a fundamental level
Design, operate, evaluate, improve
P di future
Predict
f
behavi
b h ior
• For a given syystem architecture
Professor C. Magee, 2010
Page 10
A Definition of Architecture from a Practi
tice Perspecti
tive
“An architecture is the conceptualization, description, and
design of a system, its components, their interfaces and
relationships with internal and external entities, as they
evolve over time.”
John W. Evans
Source: “Design and Inventive Engineering” Tomasz Arciszewski Fall
2004
• Similar to: “An architecture is a plan for changge.”
Joel Moses
Professor C. Magee, 2010
Page 11
Two definitions of Architecture from a Fundamental Perspective
• The architecture of a complex system is a description of the
structure or reg
gularityy of the interactions of the elements of
that system (inherently the non-random and longer lived
aspects of the system relationships).
• The architecture of a comp
plex syystem describes the functional
character of the elements and the structure of the relationships
among the elements
Professor C. Magee, 2010
Page 12
Baldwin and Clark (intermediate between practice and
fundamental?)
fundamental?)
• An architecture declares the modules and defines their
functions*
• It also declares and defines the interfaces,
interfaces including which
modules they relate and what relations are supported*
• Finally it declares or embraces standards that define common
rules of design
design, structure,
structure interfaces,
interfaces or behavior not
otherwise declared, including performance evaluation
• * are part of typical system engineering
Professor C. Magee, 2010
Page 13
Our Viewpoints
• Importance
p
of ideology
gy in framin g generic
g
architectures
and
attitudes toward them
• Importance of data and domain knowledge
• Value of doing
g case studies with quantitative results
• We will learn more about such architecture/structure by
examining a wide variety of systems such as biological,
sociological, economic at a variety of levels in addition to the
t h logiicall andd organiizational
technol
ti l systtems off most direct
di t int
i terestt to
t
us, because
• These systems are similar in many ways, perhaps more than we
think
• We will use network methods - a deliberately chosen level of
abstraction- but will also touch upon agent based models and
evolutionary dynamics
• Since we want to influence structure (not just accept it as we are
interested in design), we will also explore how structure is
determined by looking at system typologies and constraints that
influence or determine the structure
Professor C. Magee, 2010
Page 14
Lecture 1 Outline
• Organizational Information about the course
• Architectural Definitions and Role in ESD
• Systems Typollogy
ysis” Preview
• Network “Analy
• Pre-work for Lecture 2
Professor C. Magee, 2010
Page 15
Systems Context Typology I
•
•
•
•
Technical Syystems
• Power-oriented (e.g., cars, aircraft, their engines, etc.)
• Information-oriented
• Physically realized: e.g., telephone network, Internet
• Non-physical:
N
h i l e.g., software,
ft
mathematical
th
ti l systems
t
(M
(Macsyma,
Mathematica)
Organizations (of humans)
• Teams
• Hierarchies
• Networks
Social/economic “systems”
• Markets
• Social Classes
• Social networks like coauthors, citation lists, e-mails, terrorists
• Behaviors: e.g., rumors, diseases, herd mentality
Biollogiicall systems
Bi
• Cells
• Animal body plans
• The process and role of evolution
• Ecologies
Professor C. Magee, 2010
Page 16
Systems Context Typology II
• Overtly designed
• Can be an
architect
• A design strategy
is possible to
to
imagine
• Products,
product families
• Cars, aiirpllanes
• Bell System
• Organizations
Centrally-
• Centrally
planned
economies
•
Partially designed
• Non-designed
Non designed systems
• Architect not common
• No architect
• Protocols and
• Respond to context
standards are crucial
ge, develop
p
• Chang
• Design
D i strat
t tegy may
• Differentiate or
not be practical
speciate
• May be designed in
• Interact
small increments
hierarchically,
• Usually grow with
synergistically,
less direction from a
exploitatively
common strategy over
• Cells, organisms,
longer times
food webs,
webs ecological
• Regional electric grids
systems
• City streets
• Friendship
• Federal highway
groupings?
system
system
• Co-author
Co author networks?
• MIT?
In all cases, the laws of physics and legacy are important influences
D
Decomposability,
bilit hierarchy
hi
h andd time
ti dependence
d
d
are also
l significant
i ifi t in
i all
ll 3 cases
Professor C. Magee, 2010
Page 17
Systems Typology : Complex Systems Functional
Classification Matrix from Magee and de Weck
Process/Operand
Matter
(M)
Energy
(E)
Information (I)
Value
(V)
Transform or Process
(1)
GE
Polycarbonate
Manufacturing
Plant
Pilgrim
Nuclear Power
Plant
Intel Pentium V
N/A
Transport or
Di t ib t (2)
Distribute
FedEx Package
Delivery
US Power Grid
System
AT&T
Telecommunica
tion Network
Intl Banking
System
Store or
House (3)
Three Gorge
Dam
Three Gorge
Dam
Boston Public
Library (T)
Banking
Systems
Exchange or Trade (4)
eBay Trading
System (T)
Energy
Markets
Reuters News
Agency (T)
NASDAQ
Trading
System (T)
Control or Regulate
(5)
Health Care
S stem of
System
France
Atomic Energy
Commission
International
Standards
Organization
US Federal
Reser e (T)
Reserve
Professor C. Magee, 2010
Page 18
Some Things Do Not Have Architectures with Internal Structure
•
•
•
•
Random Networks
Perfect gases
Crowds of people
Their behavior can still be analyzed –indeed they are usually
easier to analyze than real systems.
systems Thus,
Thus they often form a
baseline for comparison to things that do have architectures
with significant structure
Professor C. Magee, 2010
Page 19
Structural Typology
• Totally regular
• Grids/crystals
• Pure Trees
• Layered trees
• Star ggraphs
p
• Deterministic
methods used
• Real things
• The ones we
d
are interested
in
• New methods or
adaptations of
existing methods
needed
• No internal structure
• Perfect gases
• Crowds of
people
• Classical
economics
i with
ih
invisible hand
• Stochastic methods
usedd
• Less regular
-“Hub and spokes”
-“Small Worlds”
-Communities
-Clusters
-Motifs
Motifs
Professor C. Magee, 2010
Page 20
Comments on Typologies:
Attributes of Effective Classification
•
Standards for Taxonomy
• Collectively Exhaustive and Mutually Exclusive
• Internally Homogeneous
• Stability
• Understandable Representation and Naming
•
None off thhe approachhes just reviiewed
d reall
lly fulfill
lfill thhese criiteriia. Interestiinglly
(more later in course), no categorizations of man made or evolved systems
have ever been found that fulfill these criteria. At least one natural system
categorization has been found that does fulfill these criteria (Mendeleyev)
andd has
h even been
b
the
h basis
b i off f
future successful
f l predictions.
di i
Professor C. Magee, 2010
Page 21
What has been going on recently in
“The New Science of Networks”?
• The Physicists and their friends have come to this area strongly
starting
gwith the papper in Nature by
y Watts and Stroggatz in 1998
• The publications started with a few per year and now have
reached 1000’s per year in various journals (plus 3 books).
Professor C. Magee, 2010
Page 22
Papers with “Complex Networks” in Title
Number of papers with the term"complex network" in the title.
5,200
4,800
4,400
4,000
3,600
4,820
3,200
2,800
3,701
2,400
2,000
2,310
1998
2,408
2,603
1999
2000
2,785
2001
Year
3,087
2002
2003
2004
Image by MIT OpenCourseWare.
Professor C. Magee, 2010
Page 23
What has been going on recently in
“The New Science of Networks”?
• The Physicists and their friends have come to this area strongly
starting
g with the papper in Nature by
y Watts and Stroggatz in 1998
• The publications started with a few per year and now have
reached 1000’s per year in various journals (plus 3 books).
• All of the effort builds upon work done by sociologists and
Operations Researchers over the preceding 40 or more years.
• Strong activities now exist at a variety of academic institutions:
• Th
The Uniiversit
ity off Michi
Mi higan
• Oxford University
• The Sante Fe Institute
• Columbia University
• Notre Dame University
• Many others
others
Professor C. Magee, 2010
Page 25
Why might We (who are interested in
design, management, behavior, etc.
of complex systems) care?
• A strong mathematical basis is being established for developing
relatively
y tractable models of larg
ge-scale comp
plex syystems
• We need more modeling tools that are useful for large-scale
systems with many elements, interactions and complex
behaviors
• Quantifiable metrics are being developed that may be of use in
predicting behavior of complex large-scale systems
in
• We need such metrics as they would be valuable in
•
designing and managing our systems
• Algorithms for extracting information from complex systems
are being developed and these can improve “observability” of
such systems.
• New visual representations for complex systems are being
devellopedd
Professor C. Magee, 2010
Page 26
An Architecture for ESD.342
Network Models of Systems
Networks are abstract
models that capture the
connectivity
ti it between
b t
nodes
d
plus node and arc properties
G h Theory
Graph
Th
Network metrics at
different levels
off abstraction
b t ti
Metrics developed
and used by different
intellectual traditions:
sociology
engineering
logistics
cond. matter phys
biology
ecology
economics
A
System functions and ilities
are strongly driven by
system structure (AKA architecture)
Network architecture exhibits itself in
body plans, links, clusters, paths,
modularity, spatial distributions,
flows,
o s, hierarchy,
e a c y, co
connectedness
ected ess
loops
Specific system types
and their functions and ilities
How to define,
define
recognize, and
measure ilities
Distinguish between
static and dynamic
systems
Basic
B
i Theoretical
Th
ti l Constructs
C
t t
complexity measures
percolation time (Little’s Law for system
constraint/cost of connection
B
Professor C. Magee, 2010
Page 27
B
Specific system types
and their functions and ilities
A
Methods for modeling
systems and
calculating
these metrics
Software and
algorithms for
analys
analy
sis and
display
Data and Examples:
software call lists
e mail archives
e-mail
archive s
patents
mech gizmos
coauthors
company directors
DSMs
Technical
power systems
info systems
processes
distributive and
infrastructure
software
Metrics
clustering
communities
hier. clustering
degree corr
g
rewiring
geodesics
betweenness
core-periphery
power law
Organizations
org structures
Biologi
l icall
Social/economic
food webs
local markets
DNA
input/output
models
cell processes
ethnography
evolution
rumors
ancestryy
diseases
predator-prey
Common Concepts:
canonical structures and their properties:
trees, layers, grids
commodity and info flow
percolation
modularity 1 and 2
y, failure modes
robustness, vulnerability
tradeoffs, esp of ilities
flexibility
scaling laws
power and info
critical phenomena
y stems
random and structured sy
feedback and loops
growth models
Example Systems and Their Fct-Structure Relationships:
organizations with different structures
Li-Alderson toyy networks with different bandwidths
Mechanical assemblies with different constraint
Social systems with clusters, rumors, diseases
Mathematica and Macsyma
Professor C. Magee, 2010
Page 28
Topics to think about for Thursday
• Terms and Definitions: Read the web-site document and
suggest
gg
other terms or differences you have with the “official
ESD definitions”.
• Biases and points of view about systems and structure-The
three faculty
ywill tr y to discuss their sep
parate biases and
prejudices which largely come from educational training
(undergraduate school most strongly?). We would like you
each to introduce yourselves to the class and give some
possible biases you have.
Professor C. Magee, 2010
Page 29
MIT OpenCourseWare
http://ocw.mit.edu
ESD.342 Network Representations of Complex Engineering Systems
Spring 2010
For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.