Transcript Modeling Complex Systems” by Diana Stirling

James Hornage - 090309
“Modeling Complex Systems”
by Diana Stirling
A Review and Discussion of the Paper

Complex Adaptive Systems
 Characteristics of Complex Adaptive Systems:
 They persist in spite of changes in the diverse
individual components
 The interaction between the components is what is
responsible for the persisitence of the system
 The system engages in adaptation or learning.
The Concept of Co-Evolution
 Kauffman (1995) argues that “A living system must
first be able to strike an internal compromise
between malleability and stability.
 Co-evolution is necessary because when an agent
changes, the environment around that agent
changes as well.
 Co-evolution concepts underlie the premise that
there is an optimum size and number of “patches”
to maintain maximum response ability.
Putting the Systems in Perspective
 Size and scope can range from neurons to social
systems
 Agent (system) is likely unaware of the extent of the
entire system
 Systems need only to provide excellent solutions as
opposed to optimal solutions
 Let not the pursuit of the perfect solution stand in
the way of a “perfectly good solution”.
The Power and Difficulty of Modeling
 The Power
 Modeling complex systems provides into problems
particularly when people have trouble visualizing
them.
 Modeling provides a way to understand the impact
of variables in a clear manner
 The Difficulty
 Disagreements in definition can arise, such as the
difference between Gell-Mann’s terminology and
Holland’s terminology for an “adaptive agent”
Modeling Tools
 Starlogo and Netlogo provide users with the
opportunity to see self-organization in a visual
model
 Both products use the Logo language, which was
developed out of MIT in order to teach the basics of
computing to children
 The main components of these models are turtles,
patches, and the observer.
Starlogo FireFlies
Starlogo Ants
Summary
 Complex adaptive systems:
 Involve agents, whose behavior is motivated largely
by reaction to the local environment
 Oscillate between order and chaos, which actually
provides for stability and flexibility
 Rely on excellent solutions rather than the optimum
solution
 Base their judgment about a solution as excellent by
creating a schemata based on activities that are
regularly identified with success.
Discussion
 What other models do you think would be
interesting based on the Ants and Fireflies models
from Starlogo?
 What are your thoughts on the concept of co-
evolution? It seems the environment depends on
the agent almost as much as the agent depends on
the environment?
 Do you have a different way that you would define
or describe the term complex adaptive systems?