Using genetic algorithms to optimize the parameters of the adaptive model of collective behavior of the predator attack

Dejan Grbec (2016) Using genetic algorithms to optimize the parameters of the adaptive model of collective behavior of the predator attack. MSc thesis.

Preview

PDF Download (2837Kb)

Abstract

Collective behaviour research focuses on behaviour properties of large groups of similar entities. The most studied collective behaviour species are ocks of birds and schools of �sh. There exist many di�erent computational models of collective behaviour, which researchers used to investigate various properties of collective behaviour like: transfer of information across the group, bene�ts of grouping (defence against predator, foraging),group decision{making process and group behaviour types. In this thesis we focused on various group behaviour types, transitions between them and their bene�ts when the prey group is exposed to predator attacks. In our study we implemented the collective behaviour computational model originally presented by Couzin et al. [doi:10.1006/jtbi.2002.3065], who discovered four di�erent be-haviour types by adjusting the radius of the orientation zone: swarming, rotating around an empty core, dynamic parallel movement and highly parallel movement. Swarming is typical for insects, rotating around an empty core (torus) is in special cases exhibited by sh schools and dynamic/highly parallel movement is primary associated with bird ocks and �sh schools. With known behaviour types and using various attack tactics we tried to verify which behaviour type is the most e�cient defence against predator attacks, and whether transitions between these behaviour types contribute to a more e�cient defence.

Actions (login required)