Date of Award

2015

Degree Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Graduate Group

Electrical & Systems Engineering

First Advisor

Roch Guerin

Abstract

Living in a networked world, human agents are increasingly connected as advances in technology facilitates the flow of information between and the availability of services to them. Through this research, we look at interacting agents in networked environments, and explore how their decisions are influenced by other people's decisions. In this context, an individual's decision may be regarding a concrete action, e.g., adoption of a product or service that is offered, or simply shape her opinion about a subject. Accordingly, we investigate two classes of such problems.

The first problem is the dynamics of service adoption in networked environments, where one user's adoption decision, influences the adoption decision of other users by affecting (positively or negatively) the benefits that they derive from the service. We consider this problem in the context of "User-Provided Connectivity", or UPC. The service offers an alternative to traditional infrastructure-based communication services by allowing users to share their "home base" connectivity with other users, thereby increasing their access to connectivity. We investigate when such services are viable, and propose a number of pricing policies of different complexities. The pricing policies exhibit differences in their ability to maximize the total welfare created by the service, and distributing the welfare between different stakeholders.

The second problem is the spread of opinions in a networked environment, where one agent's opinion about an issue, influences and is influenced by that of other agents to whom she is connected. We are particularly interested in the role that people's adherence to specific groups or parties may play in how final opinions are formed. We approach this problem using a model of interactions inspired by the Ising spin-glass model from classical Physics. We consider two related but distinct settings, and show that when party memberships directly influence user interactions, even slightest statistical partisan biases result in partisan final outcomes: where everyone in a party shares the same opinion, opposite to that of the other party. On the other hand, if party membership plays an indirect role in biasing agent interactions, then there is room for intra-party heterogeneity of opinions.

Files over 3MB may be slow to open. For best results, right-click and select "save as..."

Included in

Engineering Commons

Share

COinS