Algorithms and Algorithmic Game Theory for Risk Mitigation in Networks

Many of the combinatorial and network problems that motivate the study of algorithms and algorithmic game theory are traditionally modeled as deterministic. In practice, there is a lot of noise and uncertainty that can preclude the use of traditional techniques for analysis for such problems, when one is interested in reliable or risk-averse solutions.

Risk has been at the forefront of research and practice in finance and economics. However, there is still a need for developing computational approaches corresponding to these risk models, as well as new risk models and solution techniques that are specific to networked systems.

In this talk, I will give a brief overview of different definitions and approaches to risk and examples of how some core algorithmic and algorithmic game theory problems such as routing and congestion games and their analysis are transformed in the presence of uncertain network delays and risk-averse users.

I will conclude with a mix of concrete and high-level research directions that aim at a systematic understanding of risk in networks and their motivation/ applications to real world problems from transportation, telecommunications and energy.