My research investigates the mathematical foundations of decision-making under uncertainty, with an emphasis on information structures, decentralized systems, and large populations of interacting agents. I study how agents learn, coordinate, compete, and optimize when information is partial, asymmetric, or distributed across a network.
My work combines tools from stochastic control, game theory, applied probability, statistical learning, operations research, and mean-field theory.
Large stochastic teams, games, and mean-field limits
Many modern systems involve large numbers of decision-makers acting under uncertainty and partial information. Examples arise in networked control, sensor networks, energy systems, financial systems, and cooperative multi-agent systems.
My work develops existence, approximation, and structural results for stochastic teams and games with many decision-makers. A central focus is understanding when large-population limits yield tractable approximations and when symmetric or exchangeable policies are optimal.
Learning and approximation of near-optimal decisions
In many decision problems, agents do not know the model, cost, or dynamics in advance. My research studies how agents can learn from data while making decisions, with emphasis on decentralized algorithms and computationally tractable approximations.
This work uses tools from statistical learning, reproducing kernel Hilbert spaces, and operator-theoretic methods to develop scalable approaches for high-dimensional stochastic systems.
Stochastic incentive design
In hierarchical decision problems, a leader or designer seeks to influence the behavior of strategic agents. My work studies incentive policies for Stackelberg games and related stochastic systems, including settings with many followers, uncertainty, robustness requirements, and data-driven learning.
Learning in large games and teams
Multi-agent learning is difficult because each learner faces a nonstationary environment generated by the actions and learning processes of others. My research studies learning algorithms for large stochastic games and teams, with emphasis on convergence, equilibrium selection, and coordination under decentralized information.