Where is the Fastest Way Ahead to Understand - PowerPoint PPT Presentation

1 / 34
About This Presentation
Title:

Where is the Fastest Way Ahead to Understand

Description:

Where is the Fastest Way Ahead. to Understand & Design. Complex Human Systems? ... call these discoveries positive 'Black Swans': you can't predict them but you ... – PowerPoint PPT presentation

Number of Views:43
Avg rating:3.0/5.0
Slides: 35
Provided by: robert461
Learn more at: http://cs.calstatela.edu
Category:

less

Transcript and Presenter's Notes

Title: Where is the Fastest Way Ahead to Understand


1
Where is the Fastest Way Ahead to Understand
Design Complex Human Systems? The
Multi-Agent-based Simulation Path
  • R. H. Weber, Sr. P.E.
  • The Aerospace Corporation
  • (310) 336-5715

2
System Operations Engineering Reference Model
Opposed Systems (1)
National Utility
National
Utility
Operations Engr.
Military Utility
Political Utility
Economic Utility

Enterprise Archit
Mission Systems Design
Military Mission
Utility
Mission Design Effects
Space Control
Air Missile Def.

Littoral Warfare
Dominant Maneuver
Air Offense
Operational Archit
Functional
National S-A-G Intell
Performance
S-A-G Comm
S-A-G Navigation
S-A-G Surveill. Recon
Missile TW/AA
Space-Based KEW
Product Archit
Performance
System Program Level
and Cost
Constellation
Availability
Life Cycle Cost
Utility
Payload
Spacecraft
Ground
Component


System Component Level
System Engineering
Archit Design
Propulsion
Payload
Processing
Comm
Power
Structure
Software
and Cost
ADACS
CDH
Thermal
Ground
Launch
Life Cycle Cost
  • Albert Wohlstetter, Theory and Opposed-Systems
    Design,
  • RAND Report D(L)-16001-1, August 1967

S-A-G means Space--Air--Ground
3
Agent-based Engineering
  • R. H. Weber
  • H/W oriented E-O syst. engineer (USAF)
  • S/W oriented MSA (Aerospace Corp)
  • Hardened 10 yr vet. of cultural battles
  • Barriers to Progress
  • Pardon me, but your obsolete ontology is
    showing
  • US culture of individuals, innovation,
    adhocracy, organized stovepipes
  • US neglect of intellectual infrastructure
  • Maladaptive effects of Cold War on
    military-industrial complex
  • Impact ? Loss of US industrial/economic
    competitiveness at macro-system level
  • EU Airbus Consortium (Catia) Japan Toyota
    (kanban, kaizen, lean/agile mfg)
  • GOOGLE experiment agent-oriented system
    engineering ? 2 of 10 hits are US
  • agent-oriented, system,
    engineering, software ? 3 of 10 are US

4
Recognizing Overcoming Cultural Barriers
  • Will Rogers
  • Its not what people dont know that hurts them,
    its what they know that aint true.
  • Implies that education of government industry
    managers deserves top priority

5
Map of Complexity S-o-S Engineering
6
Multi-Scale Scenario Builder
  • An intentional agent at the smallest tactical or
    individual level may perpetrate actions that are
    intended to have the major effect at the
    strategic scale. The two WTC attacks eight years
    apart illustrate the difference of the first
    tactical scale event having only a regional
    effect and the second a global effect. More
    commonly it takes an integration of smaller scale
    effects within some moving time window to cause
    an effect at a larger scale. In addition to
    effects between different scales, there are
    possibilities of effects between two players
    within the same scale. Both the inter and
    intra-level effects can be described within the
    DoD paradigm of Effects Based Operations. Some
    examples of these varieties are given below.
    (Strategic S, TheaterTh, Tactical Ta)
  •  
  • S ? Ta
  • --Constrain movement check ident. at borders or
    checkpoints on roads
  • --Media reports that inhibit or encourage
    individual violence in public
  • S ? Th
  • --Failure of a leaders action encourages
    factions or coup attempts
  • --National oppression of ethnic or religious
    groups generates resistance
  • S ? S
  • --National leader stimulates change in leader of
    rival nation regarding policy risk or priority of
    resource allocation
  • --Nations negotiate for trade leverage (OPEC) or
    economic embargo (UN towards Saddam)
  • Th ? Ta
  • --Affiliation group leaders incite followers to
    riot or attack other groups
  • --Success of black market economy encourages size
    and number of gangs
  • --Coordinated attacks on utility networks
    increases disrespect for state
  • Ta ? Th
  • --Siphoning off resources (oil) from state
    increases size of black market
  • Ta ? Ta
  • --Witnesses to violence have agent state change
    from policy of avoid enemy to revenge

7
Why Agent-Oriented MS Needed for Design?
  • a priori argument its required for conceptual
    modeling
  • When complex system being designed involves
    humans in the loop for operations or processes
    involving that system (intention, goal conflicts,
    human/soft factors cognitive limitations)
  • When complex system has autocatalytic or
    autonomous control subsystems with discrete,
    multi-modal adaptive responses to environment
    (hybrid control theory, behavior of ecology and
    of adaptive life forms)
  • When complex physical phenomena involve moderate
    number of heterogeneous objects and symmetry
    breaking or phase change boundaries (eg. far from
    equilibrium condensed matter physics)
  • a posteriori argumentcase history of successful
    design impact
  • Comm computer network management via
    distributed control
  • Designing protocols for use of anti-biotics

In such cases equation-based models dont pass
face validity!
8
MS Needs Agents That are Discrete Autocatalytic
  • .The discrete character of the individuals
    turns out to be crucial for the macroscopic
    behavior of complex systems.the slightest
    microscopic granularity insures the emergence of
    localized macroscopic collective objects with
    adaptive properties.
  • The exact mechanism by which this happens depends
    crucially on the other unifying concept appearing
    ubiquitously in complex systems
    auto-catalyticity. The dynamics of a quantity is
    said auto-catalytic if the time variations of
    that quantity are proportional (via stochastic
    factors) to its current value. It turns out that
    as a rule, the "simple" objects responsible for
    the emergence of most of the complex collective
    objects in nature have auto-catalytic properties.
    Autocatalyticity insures that the behaviour of
    the entire system is dominated by the elements
    with the highest auto-catalytic growth rate
    rather than by the typical or average element.
    This has profound implications on the very
    concept of scientific explanation the fact that
    the dynamics is dominated by the exceptional
    individual/events (that enjoyed fortuitously the
    fastest stochastic growth factor) invalidates
    "reasonable" arguments based on the 'average'
    or 'representative' case. This in turn generates
    the conceptual gap separating...disciplines in
    conditions in which only a few exceptional
    individuals dominate in the emergence of nuclei
    from nucleons, molecules from atoms, DNA from
    simple molecules, humans from apes, there are
    always the un-typical casesthat carry the day.
  • This is the challenge of complexity
    understanding the basic objects (e.g. cells) in
    one science (biology) in terms of the collective
    dynamics of objects (molecules) belonging to
    another science (chemistry). Moreover the mandate
    of complexity is to uncover the determinism that
    hides behind the systematic and fateful
    recurrence in various sciences of seemingly
    fortuitous autocatalytic accidents. The
    conceptual and practical rewards for such a
    trans-disciplinary effort are inestimable."

Ref http//www.giacs.org/expertreport3
9
Roughly Three Regimes of Problems
Law of Large Numbers e ( n ) 1/2
MACRO
2) Unorganized Complexity (Aggregates)
MESO
3) Organized Complexity
(Systems)
Quantity of Objects
MICRO
Law of Medium Numbers is Murphys Law
Randomness
1) Organized Simplicity (Machines)
Combinatoric Exponential Explosion
Complexity
How many Types of Objects Interactions?
From G.M. Weinberg, An Introduction to General
Systems Thinking, John Wiley Sons, New York,
1975, p 18.
10
Matching Analysis to Types of MS
From G.M. Weinberg, An Introduction to General
Systems Thinking, John Wiley Sons, New York,
1975, p 18.
11
Limits of Equations More Specifics
  • ...there are also a number of quite concrete
    limitations to mathematical representation.
  • The difficulties of such a representation fall
    into two complementary classes those caused by
    an unrealistic treatment of time and those
    resulting from an attempt to represent multiple
    agency as an ordered sequence of individual
    actions. These difficulties are complementary
    because the unrealistic treatment of time is both
    a consequence and a partial cause of the
    unrealistic treatment of multiple agency.
  • The modeler who arranges an equation system to
    guarantee its solubility does so because he or
    she must solve it sequentially, it is not
    feasible for certain processes to be carried out
    in the background'' or for the actions of
    several agents to be revised at once. Thus only
    one agent can act at a time in such models.
    Everyone else must freeze while this action is
    taking place. The richness of the environment is
    thus restricted to suit the attention of the
    modeler. This is plainly unrealistic.
  • Edmund Chattoe , Why Are We Simulating Anyway?
    Some Answers from Economics,
  • ESRC Project L122-251-013, Nov 95

http//www.sociology.ox.ac.uk/people/chattoe.html
12
Equation-based Models in Social Sciences are
frequently the tools of charlatans.
  • in economics, and the social sciences,
    engineering has been the science of misplaced and
    misdirected concreteness. Perhaps old J.M.
    Keynes had the insight of the problem when he
    wrote To convert a model into a quantitative
    formula is to destroy its usefulness as an
    instrument of thought.
  • .Marshall, Allais and Coase used the term
    charlatanism to describe the concealment of a
    poor understanding of economics with mathematical
    smoke. Philosophers of science used the
    designation charlatanism in a the context of a
    theory that does not lend itself to falsification
    (Popper) or gradual corroboration (the
    Bayesians).
  • Against Value-at-Risk Nassim Taleb Replies to
    Philippe Jorion, 1997.

http//www.fooledbyrandomness.com/jorion.html
13
Interactive Models more Powerful than Algorithmic
  • The irreducibility of object behavior to that of
    algorithms has radical consequences for both the
    theory and the practice of computing.
  • The negative result that interaction cannot be
    modeled by algorithms leads to positive
    principles of interactive modeling by interface
    constraints that support partial descriptions of
    interactive systems whose complete behavior is
    inherently unspecifiable. The unspecifiability
    of complete behavior for interactive systems is a
    computational analog of Goedel incompleteness for
    the integers.
  • "Incompleteness is a key to expressing richer
    behavior shared by empirical models of physics
    and the natural sciences. Interaction machines
    have the behavioral power of empirical systems,
    providing a precise characterization of empirical
    computer science.
  • Peter Wegner, OOPSLA'95 Tutorial

http//www.cs.brown.edu/people/pw/
14
Control Theory View of Conflict
15
Math Modeling Resources (J. Doyle)
  • Networks of distributed sensing, computation,
    comms, and actuation will depend on all
  • Thermodynamics (Carnot)
  • Communications (Shannon)
  • Control (Bode)
  • Computation (Turing/Gödel,
  • Of these, only control addresses dynamics,
    latency, and real-time issues
  • Claim control must be the foundation for any
    network capacity theory that deals with real time
  • Focus initially on integrating comms and controls

Rhw amendments
P. Wegner)
  • Cognition (Peirce, von Foerster, Kahneman,
    Schelling, Taleb, Hawkins, )

--True only if cognition is regarded a subset of
Control
16
RobustnessFragility Trade-off (J. Doyle)
17
Critique of NRC Study on Defense M,S,A
  • Constructive--
  • Recommendation 4 DoD should establish a
    comprehensive and systematic approach for
    developing the MSA capabilities to represent
    network-centric operations
  • Enhance and sustain collaborations among the
    various parties developing network-centric MSA
    capabilities
  • the committee found little evidence of
    significant interaction and cross-fertilization
    across the application communities
    .collaboration might be facilitated by a
    DoD-sponsored series of workshopsleading to
    areport synthesizing the views of the different
    communities and identifying opportunities for
    cross-fertilization.
  • Continue and extend the development of existing
    approaches to modeling network-centric operation.
  • Since the basic architecture and functioning of
    traditional models reflect a pre-network
    perspective on military operations, those models
    are not adequate. Attention should be given to
    the use of complex agents with sizable rule sets
    governing behavior to provide quantitative models
    and to the continued coupling of agent-based
    models with the techniques of dynamic network
    analysis.
  • Establish a new mathematical basis for models
    describing network-centric operation, drawing on
    an array of approaches, particularly complex,
    adaptive systems research.

18
Critique of NRC Study on Defense M,S,A
  • Misguided Subject to Misinterpretation
  • Exploratory analysis is arguably best
    accomplished with a good aggregate-level model
    that can cover the entire possibility space
    clearly, albeit at low resolution. Such a model
    might have 6 to 10 variables.If one does such a
    synoptic exploration and finds that only two or
    three of the variables are particularly
    important, then with MRM or a suitable family of
    models, one can zoom to higher resolution on
    those variables.

19
How Will M-ABMs Improve System Design?
  • Ability to address apply Wohlstetters Opposed
    System Design Theory for systems with goal
    conflicts in quantitative simulation
  • Explicitly represent C2 system design policy
    (CONOPS) factors integrate with physical system
    engineering
  • Factor in aspects of near real-time situation
    awareness in context of scenarios with
    asymmetric, adaptive opponents
  • Allows Exploratory Analysis as a form of
    stochastic engineering (see N. Taleb in The
    Edge World Question Center) to produce more
    sustainable/adaptable systems address the
    Robust yet Fragile conundrum at lower cost by
    showing how far adaptive C2 will allow relaxing
    constraints or MOP levels on other high-cost
    system elements (eg., comm bandwidth)

20
Nassim Talebs Vision for Stochastic Science
  • Rigorous reasoning applies less to the planning
    than to the selection of what works. I also call
    these discoveries positive "Black Swans" you
    can't predict them but you know where they can
    come from and you know how they will affect you.
    My optimism in these domains comes from both the
    continuous increase in the rate of trial and
    error and the increase in uncertainty and general
    unpredictability.
  • The world is giving us more "cheap options", and
    options benefit principally from uncertainty.
    But if the success rate is very low, the more we
    search, the more likely we are to find things "by
    accident", outside the original plan or the
    more an unspecified original "plan" is likely to
    succeed. I see the sign of fractal randomness in
    these payoffs from the fact that results are more
    linear to the number of investments than they are
    to quantities invested thus favoring the
    multiplication of small bets.
  • All the while institutional science is largely
    driven by causal certainties, or the illusion of
    the ability to grasp these certainties
    stochastic tinkering does not have easy
    acceptance. Yet we are increasingly learning to
    practice it without knowing thanks to
    overconfident entrepreneurs, naive investors,
    greedy investment bankers, and aggressive venture
    capitalists brought together by the free-market
    system. I am also optimistic that the academy is
    losing its power and ability to put knowledge in
    straightjackets and more out-of-the-box knowledge
    will be generated Wiki-style.

Nassim Taleb, The Birth of Stochastic Science,
in The Edge World Question Center,
2007 http//edge.org/q2007/q07_5.htmltaleb
21
What this Workshop Can Do
  • Microscale
  • Instigate new interactions persistent
    collaboration among individuals
  • Mesoscale
  • Provide new guidance for Community of Practice
  • Share materials to educate management on
    Complexity Science via website
  • Propose new collaborative RD projects (eg. )
  • Plan successor workshop events with narrower
    focus that varies annually
  • Macroscale
  • Platform for organizing response to Europes
    GIACS (General Integration of the Applications of
    Complexity in Science) growing a global
    collaboration language for design of Complex
    Systems
  • Propose professional society recommendation for
    Chief Simulation Officer of each engineering
    corporation above 1000 employees
  • Follow-up on the more compelling recommendations
    of NRC Report Defense Modeling, Simulation
    Analysis Meeting the Challenge

22
Second Mover Contribution
  • First movers generally sacrifice peripheral
    vision in favor of focus drive
  • Building intellectual capital infrastructure
    for architecture design involves collaboration
    integration of best of breed concepts
    language that gain dominant mindshare of the
    technical community (eg., why we use Leibnitz
    notation for calculus rather than Newtons VHS
    rather than Betamax format for video tape)
  • Those who follow also serve

23
Backups
24
Overcoming Fear of Modeling 3 Stages
Each has own undocumented or unconscious
mental models - Concept design arguments
based on often conflicting assumptions
no context for resolving conflict Each
has own software models - Quantitative outputs
based on assumptions algorithms invisible to
all but model developer no std
models--little context for resolving conflict
Analysts Decisionmakers use common software
models - Shared experience of running models
with assumptions, algorithms data bases visible
to all
Barrier 1
Barrier 2
25
Co-evolution (re Brooks Turing Lecture-99)
  • Model of co-evolution from Maher Cross
  • The effective problem space evolves as the
    solution space evolves by being explored.

P1
P2
PROBLEM THREAD
S2
S1
SOLUTION THREAD
26
Outside-In Mental Modeling
  • By far the most common way to deal with something
    new is by trying to relate the novelty to what is
    familiar we think in terms of analogies and
    metaphors.
  • The only feasible way of coming to grips with
    really radical novelty is orthogonal to the
    common way of understanding it consists in
    consciously trying not to relate the phenomenon
    to what is familiar from ones accidental past,
    but approach it with a blank mind and to
    appreciate it for its internal structure.
  • The latter way of understanding is far less
    popular that the former one, as it requires hard
    thinking. (And as Bertrand Russell has pointed
    out, Many people would sooner die than thinkin
    fact they do.) It is beyond the abilities of
    thoseand they form the majorityfor whom
    continuous evolution is the only paradigm of
    history unable to cope with discontinuity, they
    cannot see it and will deny it when faced with
    it.
  • Edsger W. Dijkstra, Mathematicians Computing
    Scientists The Cultural Gap, ABACUS, vol. 4 no.
    4, Summer 1987.

27
Interactionist Approach to Architecture Design
  • When computer chips outnumber humans on this
    earththeir mediation can
  • fundamentally alter how people interact.
    Engineers, psychologists, ethnographers,
    architects, and cultural geographers have only
    begun to grasp the consequences of all this
    mediation.Much of what has passed for design has
    been an unconstrained accumulation of features,
    or at best, interfaces for measurable first-time
    usability. The new field of interaction design
    raises this work to a cultural level. As the
    study not only of how people deal with
    technology, but also how people deal with each
    other through technology, interaction design
    brings notions of premise, appropriateness, and
    appreciation to the conception of digital
    systems. The more that pervasive computing
    challenges designers to bring such notions to
    physical contexts, the more interaction design
    shares with architecture. pervasive computing
    challenges us to re-express all that we value
    most about embodiment in persistent
    structures.Now architecture incorporates
    interactivity and increasingly, interaction
    design affects architectural experience.
  • Malcolm McCullough, Visiting Associate Professor,
    School of Architecture School of Design,
    Carnegie Mellon University
  • http//www-personal.umich.edu/mmmc/

Interactionist approach to Design is another
trend supporting multi-agent based simulation
28
a priori vs a posteriori Aggregation
  • System Dynamics Models (SDM)--mature, best for
    pure physics with homogeneous elements
  • Uses equations which represent observables
    averaged over time space
  • Architecture follows equations simplest way to
    model flow rates levels
  • No explicit model of spatial relationships with
    ODE, can do with PDEs but then no way to
    differentiate between physical space network
    topology
  • Not good for behavioral discontinuities
  • Assumes homogeneity at the entity level
  • Single level of aggregated detail validation,
    no generative or atomic behavior
  • Multi Agent-Based Model (MA-BM)--new, best for
    cases of heterogeneous elements and human
    C2/policy issues
  • Begins with object/agent behavior rules governing
    interactions and aggregate observables emerge
    (multi-resolution model)
  • Natural modularity follows the types of objects
    (real world analog)
  • Can distinguish between physical space
    interaction topology
  • Handles large heterogeneity of objects
  • Behavioral validation at both object and
    aggregate levels

Refs
http//www.erim.org/cec/projects/dasch.htm
29
Attributes of Major MS Types
System Dynamics Models (SDM/EBM)
--Macro/aggregate observables generated by
equations
Multi-Agent-Based Models (M-ABM) --Agent states
change via local interaction rules (
includes but not limited to equations )
NON-Isomorphic Model
Partially Isomorphic Model
Population of AGENTS Interaction Rules
MICRO INPUT
MACRO INPUT
Equations
MACRO OUTPUT
aggregate data (m1 , m2 , )t1 , t2 ,
Output agent histories (s1 , s2 , )t1 , t2 ,
MACRO OUTPUT
Data Analysis of Aggregate Observables limited to
theory implied by generative equations
Analyst aggregates micro-level, Agent Observables
into macro-level Populations
Descriptive only--does not allow emergent
effects nor help understand their causal mechanism
Possible to discover of macro-level Effects which
can be explained by micro-level Causes
Parunak, Savit Riolo. Agent-Based Modeling vs.
Equation-Based Modeling A Case Study and Users
Guide, presented at Modeling Agent Based
Systems, 1998.
http//www.erim.org/vparunak/papers.htm
30
Simulations as Generative or Descriptive
  • Instead of being restricted to representing
    mathematical models of social processes,
  • there is no reason why simulation should not
    enable us to represent the processes themselves.
  • It seems appropriate to refer to simulations of
    this sort as generative and contrast them with
    the process of instrumental simulation discussed
    at the beginning of this section.
  • Edmund Chattoe , Why Are We Simulating Anyway?
    Some Answers from Economics, ESRC Project
    L122-251-013, Nov 95

Multi- Resolution
31
Science Limited by Infrastructure of Eq.s
  • Q Whats the Story behind this new kind of
    science?
  • A Around 1980, I had become interested in
    several really different questionsgalaxy
    formation and how brains work.the real problem
    was with the basic infrastructure of science.
    For about 300 years, most of science has been
    dominated byusing mathematical equations to
    model nature. That worked really well for
    Newtonbut its never really worked with more
    complicated phenomena in physics.in biology its
    been pretty hopeless.
  • Q If equations arent the right infrastructure
    for modeling the world, what is?
  • A Simple programs.systems in nature had
    better follow definite rules. But why should
    those rules be based on the constructs of human
    mathematics?.now you can think of them as being
    like computer programs.
  • Stephen Wolfram, Interview in New Scientist.com

http//www.newscientist.com/opinion/opinterview.js
p?idns230516
32
Open questions (J. Doyle)
Nonlinear/uncertain hybrid/stochastic etc.
Complex networked systems
?
Complexity of dynamics
Single Agent
?
Flocking/synchronization consensus
Multi-agent systems
Complexity of interconnection
33
Control Model Integrates Bode Shannon (J. Doyle)
Should also include data fusion/cognition--rhw
Nuno C Martins and Munther A Dahleh, Feedback
Control in the Presence of Noisy Channels
Bode-Like Fundamental Limitations of
Performance. (Submitted to the IEEE Transactions
on Automatic Control) Abridged version in ACC
2005 Fundamental Limitations of Disturbance
Attenuation in the Presence of Side
Information Nuno C. Martins, Munther A. Dahleh
and John C. Doyle (Submitted to the IEEE
Transactions on Automatic Control) Abridged
version in CDC 2005
http//www.glue.umd.edu/nmartins/
34
Van Riper LGen (ret) Msg to CJCS Dec 2005
  • "Systems can be complex based on the numbers of
    elements they have the greater the number of
    elements, the greater the complexity.  This is
    structural complexity.  Systems can also be
    complex in the ways that their elements
    interact  the greater the degrees of freedom of
    each element, the greater the complexity.Of the
    two, the latter can generate greater levels of
    complexity -- by orders of magnitude.
  • Within interactively complex systems it is
    usually extremely difficult, if not impossible,
    to isolate individual cases and their effects.
  • Reductive analysis will not work with such
    systems  the very act of decomposing the system
    changes the dynamics of the system
  • Most social systems, such as economies,
    governments, diplomacy, culture, and war, exhibit
    rich interactive complexity."
Write a Comment
User Comments (0)
About PowerShow.com