PostGraduate Course on Complex Systems

1
AS-74.330Post-Graduate Course on Control
EngineeringSpring 2003Complex Systems
2
AS-74.330

• Post-graduate course of the
  HUT Control Engineering Laboratory
• Changing themes, can be taken many times
• All material/presentations in English
• 6 credits out of 25-30 (major) or 10-15 (minor)

3
This is HOT!
4
Complex systems?

• Starting point (here) Simple basic system
  structure, complexity resulting from massive
  iteration
• Keyword is EMERGENCE of ORDER
• When a large number of lower-level constructs are
  combined, new patterns of functionality or
  structures (hopefully) pop up

5
System complex
System complex
Behavior chaotic
Behavior chaotic
Result fractal
Result fractal
6
Claims

• All natural processes are best explained in
  terms of cellular automata
• All natural systems are cellular automata
• The same ideas solve the problems in biology,
  sociology, ...
• Old mathematics (physics, biology, ethics, etc.)
  must be forgotten

WHAT ...?!
7
Form vs. function?
Interesting emergent forms?
8
Course objective

• Try to see the big picture.
• Be prepared to contribute to science in the
  making.
• Possibilities and threats
• Basic paradigms and approaches
• Concrete environments for emergence
• Example applications.

9

• Course pages

http//www.control.hut.fi/kurssit/AS-74.330/index.
html
10

• Read the news!

11
Use Internet!

• Links accessible through course pages
• Complex systems research
• General bookmarks
• Related journals
• Research institutions
• Information on LaTeX
• Search machines

12
Publication

• The texts will be collected together and they
  will be published in the Control Engineering
  Laboratory publication series
• Texts will be written using LaTeX, following the
  formatting guidelines
• ... an exercise in scientific writing!

13
Session outline

• Exercises to be returned before session start
• Preliminary versions of the Chapter handed out
  (can be copied in the Lab if in time)
• Some two-hour presentations, slides with
  PowerPoint, if possible
• Discussion
• Feedback forms returned.

14
Session 2

• January 22. New Kind of ScienceReino
  VirrankoskiMaterial (for example!) from Stephen
  Wolfram "A New Kind of Science" and Steven
  Johnson "Emergence - The Connected Lives of
  Ants, Brains, Cities, and Software". The goal is
  to informally give insight in the huge promises
  ...

15
Session 3

• January 29. ... or End of Science?Heikki
  PistoMaterial from John Horgan "End of
  Science". This shows the highly controversial
  nature of the claims concerning complex systems -
  and it also criticizes the "ironic science" in
  general!

16
Session 4

• February 5. Architecture of Complex
  SystemsEugene M. BurmakinMaterial from Herbert
  A. Simon "Sciences of the Artificial". The role
  of hierarchies as general structures spanning any
  complex systems is presented, and motivations for
  this fact is given ideas of "empty world" and
  "almost decomposability" are coined.

17
Session 5

• February 12. Towards DecentralizationMatti
  SaastamoinenMaterial from Stuart J. Russell,
  Peter Norvig "Artificial Intelligence - Modern
  Approach", etc. Agents and their interactions as
  a framework for mastering complexity in current
  AI paradigms is presented, resulting in extreme
  decentralization of processing and control.

18
Session 6

• February 19. Networks of AgentsJani
  KaartinenMaterial from Albert-Lazslo Barabasi
  "Linked". How everything is networked rather than
  hierarchically organized.

19
Session 7

• February 26. Cellular AutomataJuan LiMaterial
  (again) from Stephen Wolfram "A New Kind of
  Science". How all higher-level organization has
  to be abandoned?

20
Session 8

• March 5. Chaos TheoryBoris KrassiMaterial from
  Robert Devaney "An Introduction to Chaotic
  Dynamical Systems", etc. Role of nonlinearity.
  Introduction to ideas of deterministic chaos,
  bifurcations, fractals, strange attractors, and
  "Feigenbaumian universality" ...

21
Session 9

• March 12. Research Based on SimulationsLasse
  ErikssonMaterial from Per Bak, Kan Chen
  "Self-Organized Criticality", John Doyle WWW
  Homepages, etc. Further innovations about phase
  transitions, edge of chaos, highly optimized
  tolerance, ...

22
Session 10

• March 19. Self-Similarity and Power LawsTiina
  KomulainenMaterial from the Web. How the "Zipf
  law" seems to govern all naturally evolved
  systems, and how such qualitative studies can be
  utilized in artificial systems.

23
Session 11

• March 26. Turing's Lure, Gödel's CurseAbouda
  AbdullaMaterial from Elwyn Berlekamp, John
  Conway, Richard Guy "Winning Ways" (Vol. 2,
  Chapter 25), etc. How it can be theoretically
  shown that simple nonlinearities can produce
  mindboggling complexity - and how this universal
  might actually collapses everything!

24
Session 12

• April 2. Control of Large-Scale SystemsJari
  HätönenMaterial by Lauri Hakkala, etc. How the
  problem of complexity used to be attacked
  applying hierarchic control, applying optimal
  control theory before anybody spoke of robustness.

25
Session 13

• April 9. Qualitative ApproachesPetteri
  KangasMaterial from John Sterman "Business
  Dynamics - Systems Thinking and Modeling for a
  Complex World". Forgetting about details and
  numbers, it can be easier to motivate the ideas
  of System Dynamics to those who are controlling
  the complex systems ...

26
Session 14

• April 16. Towards a Systemic View of
  Complexity?Yuba Raj AdhikariMaterial from the
  Web. How emergent patterns could be controlled?
  How to make something interesting emerge? How to
  see emergent processes in a perspective? What
  kind of "holistic" mathematical tools might
  be available when analyzing complex automation
  systems?

27
Goal Homogeneity

• Presentations (oral/textual) should not be too
  superficial, and not too mathematical
• Try to see wider perspectives, historical
  developments and connections, etc.
• Give representative examples!

(It goes without saying Quality is another
central goal)
28
YOU make it roll!

• WHO are you?
• WHERE do you come from?
• WHY are you taking this course?
• WHAT do you already know about the subject?

29
Heikki Hyötyniemi

• Chairman of the Finnish Artificial Intelligence
  Society (FAIS) 1999 2002
• About 100 scientific publications
• Professor since Nov. 1, 2001 on Complex Systems