Register for: WIMSIG2017

Details of talk

TitleThe dynamics of selfish herds
PresenterShannon Dee Algar (The University Of Western Australia)
Author(s)Shannon Dee Algar
SessionDynamical Systems and Fluid Dynamics
Time15:00:00 2017-09-25
Abstract


Many well-known models of collective motion focus on the asymptotic behaviour of
groups.
Such results are not accurate representations of non-equilibrium systems where
the short time
scale motion produces the most interesting dynamics. These same models tend to
incorporate
some form of mutually compatible goal or mimicking of neighbours eventually
concluding
that a global order can emerge for an appropriate choice of parameters. We
highlight several
situations where dynamic behaviour is dominated by the presence of predators
leading one to
consider that self interest, and not cooperation, is the key driver for the
pattern formation. We
build on the notion of a selfish herd, first proposed by Hamilton in 1971,
using an interaction
network defined by the Delaunay triangulation and inertial social forces that
aim to minimise
the individualís Voronoi cell, which is used as a proxy for each individual's
positional danger.
Numerical simulations of self propelled particles illustrate that the
biologically motivated
selfish avoidance of a predator can lead to realistic and seemingly cooperative
motion
resembling that of a swarm of midges. Allowing the particles to foresee future
configurations,
a pseudo-intelligence, transforms a swarm into a flock.