Bio

Dr. Alexander Outkin is a principal member of technical staff at the Sandia National Laboratories. He has done original work in game-theoretic, mathematical and agent-based modeling in cyber security, moving target defense, financial markets, energy and economics. Alexander studied theoretical physics at Moscow State University and the Russian Academy of Sciences before receiving his PhD in economics with a specialization in game theory from Virginia Tech. He has worked as a technical staff member at Los Alamos National Laboratory. His earlier work in agent-based and microstructure modeling of the NASDAQ stock market and of the individual market-making strategies resulted in successfully predicting fundamental changes to the market structure and strategies ahead of decimalization, as well as in a book written with Vince Darley, titled A NASDAQ Market Simulation: Insights on a Major Market from the Science of Complex Adaptive Systems, published by World Scientific Publishing in 2007.

Computational Essay

Teaching Game Theory to Kids and Limits of Prediction »

Project: Algorithm Development for a Maximally Stable, Time-Dependent Graph Layout

Goal

Develop foundations for understanding, layout and display of time-dependent graphs. Preserve or develop the ability to understand the graph and the changes to it across the timeline. Take into account discontinuous graph structure or node or edge changes.

Main Results in Detail

• 1. Developed a prototype algorithm for 3D representation of time-dependent graphs evolving according to a rule from A New Kind of Science (Wolfram, 2002, p. 513).
• 2. Developed a general maximally stable, time-dependent graph layout problem and algorithm outline for graphs with different evolution rules reflecting different science or application areas. A standard 2D layout is limited in how much it helps understanding the changes, in particular across different graph iterations. The following operations were performed to enable informative 3D display: transform vertex coordinates based on vertex relative position in an individual graph and the evolving graphs timeline, center the graphs, create edges between ancestor and successor nodes and restore coordinates.

Future Work

• 1. Improve 3D graph “untangling.”
• 2. Incorporate the evolution rules explicitly into the 2D and 3D generation and display.
• 3. Enable context-specific aggregation and disaggregation.
• 4. Enable virtual reality display.
• 5. Represent history and future changes in extended-form games.