Efficient Hierarchical Approximations for Partially Observed Stochastic Optimal Control with Applications

KAUST Competitive Research Grants Program

This project develops a scalable framework for decision-making under uncertainty in partially observed systems, with applications in oceanography and environmental modeling. By integrating advanced stochastic optimization, hierarchical approximations, and Hamiltonian-based methods, it enables efficient, high-dimensional control while maintaining strong theoretical guarantees. The framework supports real-world applications such as oil-spill response, maritime routing, and energy operations.