In today’s constantly evolving world, intelligent systems must be able to make high-quality decisions in real-time in order to safely react to changing conditions and unexpected disruptions. Mathematical optimization offers a powerful and flexible tool to formulate and solve decision-making problems. Unfortunately, applying optimization to real-time decision-making has always been limited by three fundamental challenges to ensure:
- Real-time algorithm convergence to the optimal solutions on embedded platforms;
- Robustness in presence of uncertainty;
- Safety in multi-agent interactions and competition.
My research focuses on decision-making in highly dynamic and uncertain environments and lies at the interface between optimization, control, and machine learning. I use data to develop tools to accelerate optimization algorithms in fast real-time scenarios, frameworks to make decisions presence of uncertainty, and machine learning techniques to model multi-agent rationality and design interventions.
I put strong emphasis on computations, by aiming at building a trade-off between quality of the proposed techniques and computational tractability. Along the way, I seek to develop open-source numerical tools to help practitioners apply my work in the real-world. I work on various applications in autonomous systems, robotics, power systems, healthcare, finance, and engineering.