Main content

Representation, Graph Evolution, and the Induction of Desired Behaviours

Show simple item record

dc.contributor.advisor Ashlock, Daniel
dc.contributor.author Barlow, LeeAnn
dc.date.accessioned 2015-07-07T19:45:47Z
dc.date.available 2015-07-07T19:45:47Z
dc.date.copyright 2015-05
dc.date.created 2015-05-19
dc.date.issued 2015-07-07
dc.identifier.uri http://hdl.handle.net/10214/8943
dc.description.abstract Networks and combinatorial graphs are commonly used in a wide variety of types of mathematical modelling to represent structures such as contact networks in epidemiology, road networks in urban planning, and scheduling conflicts. Many of these models also use games – such as Iterated Prisoner’s Dilemma (IPD) – played such that the graph limits interaction to those who are connected by the graph to simulate more realistic social interactions. In this dissertation, we examine the novel problem of graph evolution and develop many foundational aspects that, to date, have not been explored. To effectively evolve graphs, we first develop four new graph representations, including one that generalizes all previous effective representations, called toggle-add-delete-swap (TADS). We also introduce four new benchmark functions on which to test the various representations. One of the primary goals of our graph evolution is to determine the specific graph parameters that affect the emergence of cooperation in populations playing IPD on a graph. We find that edge density is primarily responsible but that for any given edge density, it is possible to generate graphs that promote higher or lower levels of cooperation – particularly for lower edge densities. A number of potential applications for this research are suggested, including its use in designing office layouts to promote cooperation in the workplace. en_US
dc.language.iso en en_US
dc.subject evolutionary computation en_US
dc.subject graph evolution en_US
dc.subject networks en_US
dc.subject Iterated Prisoner's Dilemma en_US
dc.subject graph-based evolutionary computation en_US
dc.title Representation, Graph Evolution, and the Induction of Desired Behaviours en_US
dc.type Thesis en_US
dc.degree.programme Mathematics and Statistics en_US
dc.degree.name Doctor of Philosophy en_US
dc.degree.department Department of Mathematics and Statistics en_US


Files in this item

Files Size Format View
Barlow_LeeAnn_201507_PhD.pdf 4.501Mb PDF View/Open

This item appears in the following Collection(s)

Show simple item record