Actions

Difference between revisions of "Axtell abs"

From Santa Fe Institute Events Wiki

 
m
 
Line 8: Line 8:
 
===Thursday, January 10, 2008===
 
===Thursday, January 10, 2008===
  
1100 - 11:40 '''Rob Axtell''' ([http://atalayainstitute.org/Robert%20Axtell.html homepage])
+
11:00 - 11:40 '''Rob Axtell''' ([http://atalayainstitute.org/Robert%20Axtell.html homepage])
  
 
''Interacting Particles vs. Interacting Agents: Where is the Cut-off?''
 
''Interacting Particles vs. Interacting Agents: Where is the Cut-off?''
  
 
Interaction is a dominant theme in complex adaptive systems, in which distributed objects with heterogeneous states interact directly with one another over networks, out of equilibrium. The interacting particle systems (IPS) paradigm within probability theory is one way to represent and theorize such interactions. But in IPS each particle is typically very simple, making this more relevant for physical, chemical and some biological systems. Alternatively, the paradigm of interactive computing, emerging within computer science, is an alternative way to conceive of such interactions, and permits the objects under study to be much more complex. Seemingly, this approach is more relevant to the ecological and social sciences, where the objects represent individuals engaged in more or less purposive behavior. But the boundary between particles and agents is not clear. Sometimes particle-type models work reasonably well for social systems and it may be the case that models in which particles have agency are useful in biology, chemistry or even physics. I shall give an example of a bottom-up, agent model, in which behavior matters a great deal, but which can also be represented as an aggregate model in which agent behavior looks particle-like, being described simply by probabilities. I will point out the features of this example that lead to the cut-off of agency, and speculate as to the existence of general principles governing such cut-offs.
 
Interaction is a dominant theme in complex adaptive systems, in which distributed objects with heterogeneous states interact directly with one another over networks, out of equilibrium. The interacting particle systems (IPS) paradigm within probability theory is one way to represent and theorize such interactions. But in IPS each particle is typically very simple, making this more relevant for physical, chemical and some biological systems. Alternatively, the paradigm of interactive computing, emerging within computer science, is an alternative way to conceive of such interactions, and permits the objects under study to be much more complex. Seemingly, this approach is more relevant to the ecological and social sciences, where the objects represent individuals engaged in more or less purposive behavior. But the boundary between particles and agents is not clear. Sometimes particle-type models work reasonably well for social systems and it may be the case that models in which particles have agency are useful in biology, chemistry or even physics. I shall give an example of a bottom-up, agent model, in which behavior matters a great deal, but which can also be represented as an aggregate model in which agent behavior looks particle-like, being described simply by probabilities. I will point out the features of this example that lead to the cut-off of agency, and speculate as to the existence of general principles governing such cut-offs.

Latest revision as of 22:32, 2 January 2008

Workshop Navigation


Is There a Physics of Society? January 10-12, 2008, Santa Fe NM

Organizers: Michelle Girvan (University of Maryland) and Aaron Clauset (Santa Fe Institute)

Thursday, January 10, 2008

11:00 - 11:40 Rob Axtell (homepage)

Interacting Particles vs. Interacting Agents: Where is the Cut-off?

Interaction is a dominant theme in complex adaptive systems, in which distributed objects with heterogeneous states interact directly with one another over networks, out of equilibrium. The interacting particle systems (IPS) paradigm within probability theory is one way to represent and theorize such interactions. But in IPS each particle is typically very simple, making this more relevant for physical, chemical and some biological systems. Alternatively, the paradigm of interactive computing, emerging within computer science, is an alternative way to conceive of such interactions, and permits the objects under study to be much more complex. Seemingly, this approach is more relevant to the ecological and social sciences, where the objects represent individuals engaged in more or less purposive behavior. But the boundary between particles and agents is not clear. Sometimes particle-type models work reasonably well for social systems and it may be the case that models in which particles have agency are useful in biology, chemistry or even physics. I shall give an example of a bottom-up, agent model, in which behavior matters a great deal, but which can also be represented as an aggregate model in which agent behavior looks particle-like, being described simply by probabilities. I will point out the features of this example that lead to the cut-off of agency, and speculate as to the existence of general principles governing such cut-offs.