[Swarmfest2007] swarmfest submission for lemasson

[Jim Anderson]

Dear Steve and SwarmFest committee

I am submitting this talk for Bertrand Lemasson.

I will act as his contact information since he is currently traveling and will assume a post doc position with my group starting July.

Best Regards

 

Also, the DePaul registration site is having problems so I have not been able to register him or reserve housing.

 

Jim Anderson

206-543-4772

 

 

 

--------------------------------------------------------------------------------------------------------------------------------

Bertrand Lemasson

Utah State University and University of Washington

 

Intermediate scales of organization in collective behavior: the role of local neighborhoods and information dispersion on emergent group dynamics

 

Abstract

Schooling behavior in fish is a self-organized process that has been shown to result in rapid communication among its constituents. This adaptive behavior contributes to individual fitness in terms of foraging success, migration patterns, mating, and reducing predation risk. Simple behavioral algorithms have been demonstrated to successfully replicate the traffic rules that can generate the degree of cohesion and synchrony observed in small to moderately sized groups. In this study, we use a particle model to investigate how information is disseminated within a group to determine if information can be accurately transmitted across groups regardless of whether it originates internally, or externally. Internal stimulations are done by  assigning information on a preferred heading to a random proportion of the group. An external stimulus is mimicked by assigning informants to either the leading or trailing edge of the group. Individuals within a group are assumed to balance their social tendencies with any information they may be given, and individual movement decisions are determined based on the average heading observed within a subject’s local neighborhood. We first corroborate recent findings that only a small proportion of informed individuals are required in order to influence the directional decisions of a group. We then demonstrate that such accurate propagation of information across a group using a simple averaging rule requires that the informants be uniformly distributed throughout the group. The accuracy of directional propagation of information across a group from front to back is higher for relatively small groups (10 or 10 individuals), but reduced in groups of 50. Accuracy is reduced in all group sizes tested if the information originates at the rear of the group. Our results indicate that new traffic rules are necessary in order to explain the rapid communication dynamics observed in natural fish schools under threat.