Models and algorithms for risk-aware stochastic games

About Research Project

A central problem in building resilient systems, be those urban transport networks, asset allocation algorithms, or bank regulation frameworks, is the optimal control of risk-benefit trade-offs. Emerging literature on risk-aware and robust optimization has made significant headway in developing theory that rigorously accounts for decision makers ’risk-preferences, however, so far the competitive interaction of risk-aware agents has not been fully modeled.

In this project, we develop risk-aware game theory, in particular solution concepts and computational algorithms, by leveraging new results in the fields of risk-aware and robust optimization. We go on to apply the theory tor relevant, real-world problems, where risk awareness and competitive incentive or risk-taking interact in ways which are not presently understood, here banking, finance, and macroeconomics.

By providing solution concepts and algorithms, we expect that the result soft this research will enable a wider adoption and better design of risk-sensitive multi-agent control models. Moreover, the case studies we consider lead to insights in to enhancing the resiliency of banking and financial systems.

Project Team