*Computational physicist and applied mathematician*

Hi, I study multi-agent systems through the methods of reinforcement learning, optimal control theory, fluid mechanics, and active matter. See projects below on these topics.

*Computational physicist and applied mathematician*

Hi, I study multi-agent systems through the methods of reinforcement learning, optimal control theory, fluid mechanics, and active matter. See projects below on these topics.

Suppose you had a collection of self-driving cars where each car operates under its own local rule or policy. What is the best individual policy (or heterogenous collection of individual policies) to optimize traffic (everyone gets to their destination in a reasonable time while avoiding traffic congestion)? This project explores this question by applying Q-learning to a multi-agent gridworld system mimicking self-driving car traffic where many agents are attempting to move across a grid to their individually prescribed destination while avoiding collision.

What is the best way to mix two fluids? Say --- cream and coffee? We explore this question by using concepts from numerical optimization, optimal control theory, and functional analysis.

The linked article below is an interactive educational piece on complex systems. It attempts to explain complex systems science by exploring two classic bird flocking models.

Swimming bacteria rarely exist in isolation. Rather, they are commonly found in groups where bacteria can interact with each other. When the main form of interaction is mediated through the fluid flow between them, interesting rich global patterns can emerge as a result. We adopt the modeling perspective that the collection of bacteria operates as a collection of active self-propelling particles. We formulate this perspective by using concepts from fluid dynamics and kinetic theory to study the collective dynamics of various initial particle distributions.

See older projects, here: