Transcript lecture9 - NYU Computer Science Department

Lecture 11
Echoing the emergence
Santa Fe Institute
• Formed in 1985
• Think tank to deal with complexity
• Scientists from all areas including
physics, chemistry, biology, computer
science, economics, ecology, sociology,
history, etc.
• http://www.santafe.edu
The Flow
• Dynamical systems
• We find ourselves among ever-changing systems
• There is an intractable number of branches
• Feedback
• Systems change and develop by receiving feedback from the
environment and responding to environment. Thus systems
are inextricable parts of the environment
• Self-reference
• The systems around us are heavily recursive, selfbootstrapping
• Co-evolution
• There is no stand alone evolution, everything is co-evolving.
• Everything is dependent and influences everything else.
Cybernetics
• Norbert Wiener
• Trying to understand how control & communication worked
• Greek kybernetes (steersman) (Web of life pp.. 97)
• We are but whirlpools in a river of ever-flowing water
• Self-regulation
• We are not stuff that abides, but patterns that perpetuate
themselves
Situation
Assessment
Action
Impact
on
environment
Ilya Prigogine
• Second law of thermodynamics
• In closed system, the amount of entropy in a given system does
not decrease
• Entropy means disorder
• Living organisms and equilibrium
• Living beings are in order, away from equilibrium
• Open thermodynamical systems
• Nonlinear equations
• Self-Organization
Equilibrium
(ice)
Order
(edge of chaos)
Chaos
(gas)
Prisoners’ Dilemma
- Two thieves are captured and both promised life if
they would tell on each other. The problem is that
ideally, they should simply just not tell on each
other. But one of they think 'What if the other guy
tells on me?' Then I will be in jail, and he will be
spared. So they both defect, i.e. tell on each other
and loose.
- Iterative nature of biological games
Seven basic elements of CAS
• Aggregation
– Economy and markets
– Body and nervous, immune, endocrine system
– World economy and country economies
–
–
–
–
Emergence as a result of interactions
whole > sum of the parts
higher level of organization
meta agents
Seven basic elements of CAS
• Tagging (mechanism)
• Identification of alike agents
• Grouping
• Attribute
– contracts between firms
– form of adaptation - delegation
• Divisions in a firm Equities, Fixed Income, etc.
– collaboration, formation of aggregate and diversification
via tagging
Seven basic elements of CAS
• Nonlinearity (property)
– aggregation & tagging
– threshold of emergence (H >sum(P))
– predator/prey interaction
• One of the standard example of nonlinear dynamical model is
predator/prey interaction. Observe that increases in either
population increase the likelihood of a contact. Let Predator(t),
and Prey(t) be number of predators and prey at some time t, and
let c be the constant that reflects efficiency of a predator. We can
calculate the number of interactions per unit of time as
c*Predator(t)*Prey(t). That is, with c = 0.5, Predator(t) = 2 and
Prey(t) = 10, we would have 10 encounters. Now, let us double
each population so that Predator(t) = 4 and Prey(t) = 20, then we
will have 40 encounters.
– nonlinearity is a result of a product instead of a sum
Seven basic elements of CAS
• Flow (property)
– nonlinearity induces flow
– multiplier effect
– feedback and cycles
= Dead
= Alive
Seven basic elements of CAS
• Diversity (property)
– arise from exploration of multitude of possibilities (local
adaptations)
– firms enter and leave market
– mimicry
Seven basic elements of CAS
• Internal models (mechanism)
– ‘anticipation’
– survival of the fittest
– subconscious mode
Seven basic elements of CAS
• Building blocks (mechanism)
– decomposition
– quark, nucleon, atom, molecule, organelle, cell
– generation of internal models
Eworld Parameters
Eworld Resources
<RESOURCE
CLASS=edu.nyu.sejava.iskold.eworld.HomogeneousResource>
<PROPERTY NAME=Quantity VALUE=25>
<PROPERTY NAME=Base VALUE=a>
<PROPERTY NAME=Renewable VALUE=true>
</RESOURCE>
Eworld Resources
<RESOURCE
CLASS=edu.nyu.sejava.iskold.eworld.HomogeneousResource
<RBASE>
<PROPERTY NAME=Quantity VALUE=2>
<PROPERTY NAME=Base VALUE=a>
</RBASE>
<RBASE>
<PROPERTY NAME=Quantity VALUE=2>
<PROPERTY NAME=Base VALUE=t>
</RBASE>
<PROPERTY NAME=Renewable VALUE=false>
</RESOURCE>
Entity’s Genome
- Reservoir (place where excess resources are stored)
- 3 PhenoGenes
(externally visible)
- OffenseGene
- Defense Gene
- Mating Gene
- 3 Rule Genes
(internal only)
- CombatConditionGene
- TradeConditionGene
- MatingConditionGene
- Resource collection gene
- Trading resource base
Entity Generators
<GENERATOR
CLASS=edu.nyu.sejava.iskold.eworld.BasicRandomEntityGenera
tor>
<PROPERTY NAME=Quantity VALUE=25>
</RESOURCE>
<GENERATOR
CLASS=edu.nyu.sejava.iskold.eworld.BasicConstrainedEntityGe
nerator>
<PROPERTY NAME=Quantity VALUE=25>
<PROPERTY NAME=OffenseGeneLength VALUE=5>
<PROPERTY NAME=OffenseGeneBases VALUE=AT>
</RESOURCE>
Entity near Resource
- Is this resource empty?
- Does my Resource Collection Gene allow me to pick up
next base from this resource?
NOPE, NO MATCH
C
AT G
NEXT BASE IN SOME RESOURCE
Cloning
- Does my Reservoir contain enough bases to clone
myself?
Cloning threshold is at least 1
Offense ACG
Defense AAA
Mating CGT
Combat Condition AAGT
Trade Condition GGG
Mating Condition CG
Resource collection AG
Trading resource A
RESERVOIR CONTAINS TWICE THE
GENOME
16 A
4T
6C
16 G
Combat
Combat
Single point crossover
Single point crossover
Mortality
- Every round kill entities at random
- The longer entity lives the higher its chance of
dying
Statistics
Some book references
Complexity,
by M. Mitchell Waldrop
At Home in the Universe,
by Stuart Kauffman
Web Of Life,
by Frijof Capra
The Blind Watchmaker,
by Richard Dawkins