Optimization, multi-agent learning & applied mathematics.
A federated multi-agent deep RL (FL-MADRL) framework that jointly optimizes UAV trajectories and reconfigurable intelligent surface phase shifts across a space–air–ground integrated network — improving coverage and spectral efficiency while keeping training privacy-preserving and decentralized.
Full publication & citation record via ORCID 0009-0006-8255-8711.
Evaluate and validate the scientific reasoning behind novel algorithms: formal problem formulation, mathematical derivation, proofs of properties and convergence analysis (MADDPG, HFL-MADRL). Author and reviewer of IEEE publications; secured competitive grants; collaborated with the Hon Hai (Foxconn) Research Institute. Thesis: multi-agent deep reinforcement learning & federated learning for intelligent spatial systems.
Translated theoretical signal models into a verified, working real-time communication prototype using Software-Defined Radio (SDR).
Developed and analyzed optimization methods for wireless networks; modeled aerial base stations (UAV-BS) and Reconfigurable Intelligent Surfaces (RIS). Published peer-reviewed research and co-led grant proposals with industry partners.
Available for research collaboration, scientific adjudication and AI evaluation roles.