Document Type

Thesis

Publication Date

2014

Advisor

Yu Zhang

Abstract

The study of social networks has become increasingly important in recent years. Multi-agent systems research has proven to be an effective way of representing both static and dynamic social networks in order to model and analyze many different situations. Previous implementations of multi-agent systems have observed a phenomenon called tolerance between agents through simulation studies, which is defined as an agent maintaining an unrewarding connection. This concept has also arisen in the social sciences through the study of networks. We aim to bridge this gap between simulation studies in multi-agent systems and real-world observations. This project explores how local interactions of autonomous agents in a network relate to the development of tolerance. We have developed a new model for multi-agent system interactions based on these observations. We also claim that tolerance is directly observable in real dynamic social networks, and the parameters that govern tolerance of a system can be estimated using a Hidden Markov Model.

Comments

Readers: Kris Nairn, Phil Byrne, Robert Hesse, Imad Rahal

Share

COinS