Decentralized Interactions In Networks
Open Access
- Author:
- Zhang, Yang
- Graduate Program:
- Business Administration
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- July 31, 2013
- Committee Members:
- Gary Bolton, Dissertation Advisor/Co-Advisor
Susan Xu, Committee Chair/Co-Chair
Elena Katok, Committee Member
Anthony Mark Kwasnica, Committee Member
Kalyan Chatterjee, Committee Member - Keywords:
- network
game theory
behavioral experiment
operations research - Abstract:
- My dissertation concerns decentralized decision making of networked agents. By means of laboratory experiments, Chapter 1 and 2 respectively examine how the network structure affects coordination and free riding, referring to the case where a player’s incentive to take a certain action increases / decreases if more players connected to her takes the same action. Typical applications include technology adoptions under network effects (coordination), and the supply of informational goods in peer-to-peer networks (free riding). In Chapter 1, we find that both global and local networks impact coordination when agents have incomplete information about the network structure. Simple topological measures (e.g. network density and one’s number of connections) appear sufficient for predicting the level of successful coordination. In Chapter 2, we find a strong effect of one’s local network on free riding, which is robust regardless of whether agents have full or localized view of the network. The patterns of free riding under both information settings can be explained by a unified behavioral model involving one’s local network connectivity, and one’s tendency of reacting to past neighbor actions and actions of oneself. In Chapter 3 we study how a network is formed by decentralized decisions of agents. Through a simple analytical model, we explain the emergence of core-periphery landscape in knowledge citation networks as a dynamic process, driven by innovation and citation of individual forward-looking authors who contribute knowledge in their own interests. We show that the core-periphery structure forms on the equilibrium path with an initial state that contains sufficiently large amount of knowledge.