36
Toward a Better Understanding of Complexity —————————————— Definitions of Complexity, Cellular Automata as Models of Complexity, Random Boolean Networks Christian Jacob [email protected] Department of Computer Science University of Calgary

Toward a Better Understanding of Complexitypages.cpsc.ucalgary.ca/.../Slides/CA-Complexity-bw.pdf · 2002. 1. 25. · Definition s o f C o mplexity F Information – The capability

  • Upload
    others

  • View
    5

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Toward a Better Understanding of Complexitypages.cpsc.ucalgary.ca/.../Slides/CA-Complexity-bw.pdf · 2002. 1. 25. · Definition s o f C o mplexity F Information – The capability

Toward a Better Understanding ofComplexity

——————————————Definitions of Complexity, Cellular Automata as

Models of Complexity, Random Boolean Networks

Christian [email protected]

Department of Computer Science

University of Calgary

Page 2: Toward a Better Understanding of Complexitypages.cpsc.ucalgary.ca/.../Slides/CA-Complexity-bw.pdf · 2002. 1. 25. · Definition s o f C o mplexity F Information – The capability

Complexity

Definitions

Page 3: Toward a Better Understanding of Complexitypages.cpsc.ucalgary.ca/.../Slides/CA-Complexity-bw.pdf · 2002. 1. 25. · Definition s o f C o mplexity F Information – The capability

Definitions of Complexity

F Information– The capability of a system to “surprise” an observer, i.e.

to provide information.

F Time-Computational Complexity– The time an algorithm needs to solve a problem.

F Space-Computational Complexity– The amount of memory an algorithm needs to solve a

problem.

Page 4: Toward a Better Understanding of Complexitypages.cpsc.ucalgary.ca/.../Slides/CA-Complexity-bw.pdf · 2002. 1. 25. · Definition s o f C o mplexity F Information – The capability

Definitions of Complexity (2)

F Effective Complexity– The degree of order (instead of randomness) of a

system.

F Entropy– The complexity of a system is equal to the thermo-

dynamical measure of its disorder.

Page 5: Toward a Better Understanding of Complexitypages.cpsc.ucalgary.ca/.../Slides/CA-Complexity-bw.pdf · 2002. 1. 25. · Definition s o f C o mplexity F Information – The capability

Definitions of Complexity (3)

F Fractal Dimension– The “fuzziness” of a system, measuring the degrees of

details a system reveals on arbitrary scales.

F Hierarchical Complexity– The diversity of different layers, which a hierarchical

system is composed of.

F Mutual Information– The degree to which a part of a system has information

about other involved system constituents.

Page 6: Toward a Better Understanding of Complexitypages.cpsc.ucalgary.ca/.../Slides/CA-Complexity-bw.pdf · 2002. 1. 25. · Definition s o f C o mplexity F Information – The capability

Definitions of Complexity (4)

F Information Distance– The differences among parts of a system.

F Grammatical Complexity– The degree of universality a language must have to

describe a system (regular, context-free, context-sensitive, universal Turing machine)

F …

Page 7: Toward a Better Understanding of Complexitypages.cpsc.ucalgary.ca/.../Slides/CA-Complexity-bw.pdf · 2002. 1. 25. · Definition s o f C o mplexity F Information – The capability

Cellular Automata

Random BooleanNetworks

Classifier Systems

SwarmSystems

7

Page 8: Toward a Better Understanding of Complexitypages.cpsc.ucalgary.ca/.../Slides/CA-Complexity-bw.pdf · 2002. 1. 25. · Definition s o f C o mplexity F Information – The capability

Cellular Automata

Global Effects from LocalRules

Page 9: Toward a Better Understanding of Complexitypages.cpsc.ucalgary.ca/.../Slides/CA-Complexity-bw.pdf · 2002. 1. 25. · Definition s o f C o mplexity F Information – The capability

Cellular Automata

F The CA space is a lattice of cells with aparticular geometry.

F Each cell contains a variable from a limitedrange (e.g., 0 and 1).

F All cells update synchronously.

F All cells use the same updating rule,depending only on local relations.

F Time advances in discrete steps.9

Page 10: Toward a Better Understanding of Complexitypages.cpsc.ucalgary.ca/.../Slides/CA-Complexity-bw.pdf · 2002. 1. 25. · Definition s o f C o mplexity F Information – The capability

One-dimensional finite CA architecture

time

F K = 5 localconnectionsper cell

F Synchronousupdate in discretetime steps

A. Wuensche: The Ghost in the Machine, Artificial Life III, 1994. 10

Page 11: Toward a Better Understanding of Complexitypages.cpsc.ucalgary.ca/.../Slides/CA-Complexity-bw.pdf · 2002. 1. 25. · Definition s o f C o mplexity F Information – The capability

Cellular Automata:Local Rules — Global Effects

11

Page 12: Toward a Better Understanding of Complexitypages.cpsc.ucalgary.ca/.../Slides/CA-Complexity-bw.pdf · 2002. 1. 25. · Definition s o f C o mplexity F Information – The capability

Time Evolution of the ith Cell

Ci(t+1) = f (Ci -[ K / 2]

( t ) ,..., Ci -1( t) ,Ci

( t ),Ci +1( t ) ,..., Ci +[ K / 2]

( t ) )

With periodic boundary conditions:

x < 1: Cx = CN+ x x > N : Cx = Cx - N

12

Page 13: Toward a Better Understanding of Complexitypages.cpsc.ucalgary.ca/.../Slides/CA-Complexity-bw.pdf · 2002. 1. 25. · Definition s o f C o mplexity F Information – The capability

Value Range and Update Rules

F For V different states (= values) per cellthere are VK permuations of values in aneighbourhood of size K.

F The update function f can be implementedas a lookup table with VK entries, givingVVK

possible rules.

13

Page 14: Toward a Better Understanding of Complexitypages.cpsc.ucalgary.ca/.../Slides/CA-Complexity-bw.pdf · 2002. 1. 25. · Definition s o f C o mplexity F Information – The capability

Example Update Rule

F V = 2, K = 3

F The rule table for rule 30:

111 110 101 100 011 010 001 000 0 0 0 1 1 1 1 0

14

Page 15: Toward a Better Understanding of Complexitypages.cpsc.ucalgary.ca/.../Slides/CA-Complexity-bw.pdf · 2002. 1. 25. · Definition s o f C o mplexity F Information – The capability

CA Demos

F Evolvica CA Notebooks

Page 16: Toward a Better Understanding of Complexitypages.cpsc.ucalgary.ca/.../Slides/CA-Complexity-bw.pdf · 2002. 1. 25. · Definition s o f C o mplexity F Information – The capability

Cellular Automata

Models of Complexity(Stephen Wolfram Approach)

Page 17: Toward a Better Understanding of Complexitypages.cpsc.ucalgary.ca/.../Slides/CA-Complexity-bw.pdf · 2002. 1. 25. · Definition s o f C o mplexity F Information – The capability

Analysis of Cellular Automata

F … as Discrete Dynamical Systems

– Discrete idealization of partial differential equations

– Set of possible (infinite) CA configurations forms aCantor set

– CA evolution may be viewed as a continuous mappingon this Cantor set

– Entropies, fractal dimensions, Lyapunov exponents

Page 18: Toward a Better Understanding of Complexitypages.cpsc.ucalgary.ca/.../Slides/CA-Complexity-bw.pdf · 2002. 1. 25. · Definition s o f C o mplexity F Information – The capability

Analysis of Cellular Automata

F … as Information-Processing Systems

– Parallel-processing computer with a simple gridarchitecture.

– Initial configuration is processed by theevolution of the cellular automaton.

– What types of formal languages are generated?

Page 19: Toward a Better Understanding of Complexitypages.cpsc.ucalgary.ca/.../Slides/CA-Complexity-bw.pdf · 2002. 1. 25. · Definition s o f C o mplexity F Information – The capability

Four Classes of Patterns

F Wolfram classifies CAs according to the patternsthey evolve:

– 1. Pattern disappears with time.– 2. Pattern evolves to a fixed finite size.– 3. Pattern grows indefinitely at a fixed speed.– 4. Pattern grows and contracts irregularly.

– 3/text.html: Fig. 1

Page 20: Toward a Better Understanding of Complexitypages.cpsc.ucalgary.ca/.../Slides/CA-Complexity-bw.pdf · 2002. 1. 25. · Definition s o f C o mplexity F Information – The capability

Four Qualitative Classes

F 1. Spatially homogeneous stateF 2. Sequence of simple stable or periodic structuresF 3. Chaotic aperiodic behaviourF 4. Complicated localized structures, some

propagating

• 85-cellular/7/text.html: Fig. 3 (first row)

Page 21: Toward a Better Understanding of Complexitypages.cpsc.ucalgary.ca/.../Slides/CA-Complexity-bw.pdf · 2002. 1. 25. · Definition s o f C o mplexity F Information – The capability

Classes from an InformationPropagation Perspective

F 1. No change in final state

F 2. Changes only in a finite region

F 3. Changes over an ever-increasing region

F 4. Irregular changes

Page 22: Toward a Better Understanding of Complexitypages.cpsc.ucalgary.ca/.../Slides/CA-Complexity-bw.pdf · 2002. 1. 25. · Definition s o f C o mplexity F Information – The capability

Degrees of Predictability for theOutcome of CA Evolution

F 1. Entirely predictable, independent ofinitial state

F 2. Local behavior predictable from localinitial state

F 3. Behavior depends on an ever-increasinginitial region

F 4. Behavior effectively unpredictable

Page 23: Toward a Better Understanding of Complexitypages.cpsc.ucalgary.ca/.../Slides/CA-Complexity-bw.pdf · 2002. 1. 25. · Definition s o f C o mplexity F Information – The capability

Suggested Explorations:

F Natural CA-like Phenomena (CBN, 15.5)F Stephen Wolfram’s vs. Chris Langton’s CA

Classification (CBN, 15.2, 15.3)F Conway’s Game of Life (CBN, 15.4)F Self-Similarity and Fractal Geometry (CBN,

Ch. 5)F L-Systems and Fractal Growth (CBN, Ch.

6)

Page 24: Toward a Better Understanding of Complexitypages.cpsc.ucalgary.ca/.../Slides/CA-Complexity-bw.pdf · 2002. 1. 25. · Definition s o f C o mplexity F Information – The capability

Suggested Explorations: (2)

F Fractals (CBN, Ch. 9):– Simplicity and Complexity

F Nonlinear Dynamics in Simple Maps– Logistic map, stability vs. instability,

bifurcations, chaos (CBN, Ch. 10)

F Strange Attractors– Hénon-Lorenz attractor (CBN, Ch. 11)

Page 25: Toward a Better Understanding of Complexitypages.cpsc.ucalgary.ca/.../Slides/CA-Complexity-bw.pdf · 2002. 1. 25. · Definition s o f C o mplexity F Information – The capability

Modeling Excitable Media

• Slime mold growth,

• star formation in spiral disk galaxies,

• cardiac tissue contraction,

• diffusion-reaction chemical systems and

• infectious disease epidemics

– would seem to be quite dissimilar systems.

– Yet, in each of these cases, various spatiallydistributed patterns, such as concentric andspiral wave patterns, are spontaneously formed.

Page 26: Toward a Better Understanding of Complexitypages.cpsc.ucalgary.ca/.../Slides/CA-Complexity-bw.pdf · 2002. 1. 25. · Definition s o f C o mplexity F Information – The capability

Modeling Excitable Media

• The underlying cause of the formation of these self-organized, self-propagating structures is that theseare excitable media, consisting of spatiallydistributed elements which can

– become excited as a result of interacting with neighboringelements,

– subsequently returning incrementally to the quiescent state

– in which they are again receptive to being excited.

• The excited-refractory-receptive cycle thatcharacterizes these systems can be modelled usingmulti-state cellular automata.

Page 27: Toward a Better Understanding of Complexitypages.cpsc.ucalgary.ca/.../Slides/CA-Complexity-bw.pdf · 2002. 1. 25. · Definition s o f C o mplexity F Information – The capability

2-D CA:Emergent Pattern Formation in Excitable Media

Neuron excitation

Neuron excitation (relaxed)

Hodgepodge

27

Page 28: Toward a Better Understanding of Complexitypages.cpsc.ucalgary.ca/.../Slides/CA-Complexity-bw.pdf · 2002. 1. 25. · Definition s o f C o mplexity F Information – The capability

Cellular Automata

Random BooleanNetworks

Classifier Systems

SwarmSystems

28

Page 29: Toward a Better Understanding of Complexitypages.cpsc.ucalgary.ca/.../Slides/CA-Complexity-bw.pdf · 2002. 1. 25. · Definition s o f C o mplexity F Information – The capability

Complex Systems

EmergentBehaviours and Patterns

fromLocal Interactions

Page 30: Toward a Better Understanding of Complexitypages.cpsc.ucalgary.ca/.../Slides/CA-Complexity-bw.pdf · 2002. 1. 25. · Definition s o f C o mplexity F Information – The capability

Stevens et al., 1988

Page 31: Toward a Better Understanding of Complexitypages.cpsc.ucalgary.ca/.../Slides/CA-Complexity-bw.pdf · 2002. 1. 25. · Definition s o f C o mplexity F Information – The capability

Nuridsany & Pérennou, 1996

Page 32: Toward a Better Understanding of Complexitypages.cpsc.ucalgary.ca/.../Slides/CA-Complexity-bw.pdf · 2002. 1. 25. · Definition s o f C o mplexity F Information – The capability

Ernst, 1998

Nuridsany & Pérennou, 1996

Page 33: Toward a Better Understanding of Complexitypages.cpsc.ucalgary.ca/.../Slides/CA-Complexity-bw.pdf · 2002. 1. 25. · Definition s o f C o mplexity F Information – The capability

What to Learn from Ant Colonies asComplex Systems

F Fairly simple units generate complicatedglobal behaviour.

F “If we knew how an ant colony works, wemight understand more about how all suchsystems work, from brains to ecosystems.”(Gordon, 1999)

Page 34: Toward a Better Understanding of Complexitypages.cpsc.ucalgary.ca/.../Slides/CA-Complexity-bw.pdf · 2002. 1. 25. · Definition s o f C o mplexity F Information – The capability

Emergence in Complex Systems

F How do neurons respond to each other in away that produces thoughts?

F How do cells respond to each other in a waythat produces the distinct tissues of agrowing embryo?

F How do species interact to producepredictable changes, over time, inecological communities?

F ...

Page 35: Toward a Better Understanding of Complexitypages.cpsc.ucalgary.ca/.../Slides/CA-Complexity-bw.pdf · 2002. 1. 25. · Definition s o f C o mplexity F Information – The capability

Complexity through EmergenceExamples of Swarm Systems ...

Page 36: Toward a Better Understanding of Complexitypages.cpsc.ucalgary.ca/.../Slides/CA-Complexity-bw.pdf · 2002. 1. 25. · Definition s o f C o mplexity F Information – The capability

References

F Bonabeau, E., Dorigo, M., and Theraulaz, G. (1999). SwarmIntelligence: From Natural to Artificial Systems. New York, OxfordUniversity Press.

F Ernst, A. M., ed. (1998). Digest: Kooperation und Konkurrenz,Heidelberg, Spektrum Akademischer Verlag.

F Gordon, D. (1999). Ants at Work. New York, The Free Press.F Hölldobler, B., and Wilson, E. O. (1990). The Ants. Cambridge,

MA, Harvard University Press.F Nuridsany, C., and Pérennou, M. (1996). Microcosmos: The

Invisible World of Insects. New York, Stewart, Tabori & Chang.F Resnik, M. (1997). Turtles, Termites, and Traffic Jams.

Cambridge, MA, MIT Press.F Stevens, C. F., et al. (1988). Gehirn und Nervensystem.

Heidelberg, Spektrum Akademischer Verlag.