My thesis focuses on models for complex networks. This work was performed in collaboration with Ed Scheinerman at Johns Hopkins University and with my advisor, Milena Mihail (College of Computing) on generalizing the random dot product graph model. In partiuclar, we develop a robust and general model for complex networks based on the dot product. We are able to show that there is a general reduction of the diameter of the random dot product graph model to the diameter of a suitable class of Erdos-Renyi graphs. Further, we show that the model is clustered in a natural way and there is an explicit relationship between the underlying geometry of the random space to the degree distribution of the resulting model. This results generalize to the natural directed graph model. We also explore other modeling and algorithmic aspects of the random dot product graph model.
I am also currently doing research with Tom Trotter and his graduate students (Csaba, Dave, and Mitch) on the combinatorics of partially ordered sets and related questions. This collaboration has resulted in partial result on the structure of forbidden subposets for linear discrepancy 3 (see paper below).
On the side I am also working with Joel Sokol from the School of Industrial and Systems Engineering on a project in choice aggregation as it relates to the professional baseball draft. In essence, since there are so many players eligble for the draft (approximately 2000), there is no way for one scout to see them all, so a team must have some way of aggregating all of its scout's limited and potentially contradictory information.
Contact Information
Stephen J. Young
Skiles Building, Room 250
686 Cherry Street
Georgia Institute of Technology
Atlanta, Georgia 30332-0160
Phone: 404-385-2468
young@math.gatech.edu