Skip navigation
Please use this identifier to cite or link to this item:
Title: An Analysis of Zero-Sum Games on Networks with Voter Model Dynamics
Authors: Aresti, Casilda
Advisors: van Handel, Ramon
Department: Operations Research and Financial Engineering
Class Year: 2015
Abstract: This paper will begin by introducing several existing opinion dynamics models and assessing their ability to represent empirical observations accurately. We will then exclusively focus on the classic voter model and consider the effects of introducing a small number of stubborn agents that can hold one of two different opinions. Stubborn agents holding the same opinion are considered to belong to the same party and to have directly opposing objectives to those belonging to the other party. Assuming that each party’s objective is to maximize its expected number of supporters as t → ∞, we will explore the optimal strategies for each party in both sequential and simultaneous games. In the sequential case, we will propose the ‘convergence time minimization’ strategy as a tractable proxy for the first mover’s optimal strategy and assess its performance on network structures with varying degrees of homogeneity. In the simultaneous case, we will seek to qualitatively understand the conditions under which a pure strategy Nash equilibrium is expected to arise. Finally, we will explore how the parties’ optimal strategies would change if the objective function were adjusted so as to incorporate variance or if it were evaluated at a finite time horizon. Our analysis and conclusions find applications in the commercial and political fields, where a small number of entities seek to diffuse opposing beliefs through the positioning of advertising resources.
Extent: 91 pages
Type of Material: Princeton University Senior Theses
Language: en_US
Appears in Collections:Operations Research and Financial Engineering, 2000-2017

Files in This Item:
File SizeFormat 
PUTheses2015-Aresti_Casilda.pdf1.03 MBAdobe PDF    Request a copy

Items in Dataspace are protected by copyright, with all rights reserved, unless otherwise indicated.