Hao Xu
University of Nevada-Reno, NV·Engineering·Early-career
Many modern systems contain huge numbers of interacting decision-makers, or agents. Examples include wide-area power management systems, smart transportation, ultra-dense 5G and 6G networks, and the wireless networks inside factories.
- Fields
Engineering
Computer Science
Mathematics & Statistics
- Topics
multi-agent systems
game theory
reinforcement learning
- Grants
- 2 known grants, through Mar 2027
What they work on
- Develop a hierarchical game theory (HGT) for large-scale multi-agent systems, such as power management and transportation networks, that maintains optimal efficiency while keeping computing demands manageable as agent numbers grow.
- Create an actor-critic reinforcement learning method based on backward stochastic differential equations to solve high-dimensional HGT optimization problems for large multi-agent systems.
- Build a quality-of-performance-driven hybrid reinforcement learning framework that balances learning efficiency and computing cost with guaranteed safety, enabling real-time optimization under uncertainty.
- Design hardware-driven cross-layer optimization for industrial wireless networks assisted by reconfigurable intelligent surfaces (RIS), solved with real-time data-enabled learning under harsh constraints and unknown disturbance.
- Develop a robust, computationally efficient learning framework that optimizes large RIS-enhanced wireless networks in a distributed way, for stationary and mobile factory users.
- Design and fabricate an RIS unit with dynamic beam steering and build a hardware testbed to evaluate RIS-enhanced industrial wireless networks in practical settings.
More about the research
Running such systems efficiently and reliably in real time, despite uncertainty and disturbances, is hard. The computing needed for optimization and learning explodes as the number of agents grows, a problem known as the curse of dimensionality. The lab's work combines game theory (the mathematics of interacting decision-makers), reinforcement learning (algorithms that learn by trial and feedback) and optimization. One project develops general theory for large multi-agent systems. A second applies learning and optimization to wireless networks for smart factories.
Why it matters
Better methods for coordinating many agents could make power management systems, smart transportation and dense 5G/6G networks more efficient and resilient under uncertainty. Reliable, secure and low-delay wireless networks built on reconfigurable intelligent surfaces could help smart factories connect their sensors and machines, supporting advanced manufacturing.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
CAREER: Toward Hierarchical Game Theory and Hybrid Learning Framework for Safe, Efficient Large-scale Multi-agent Systems
NSF · Apr 2022 – Mar 2027
Collaborative Research: SWIFT: Data Driven Learning and Optimization in Reconfigurable Intelligent Surface Enabled Industrial Wireless Network for Advanced Manufacturing
NSF · Oct 2021 – Sep 2026
Closest labs
Zejian Zhou
University of Wyoming, WY·Engineering
Model complex autonomous systems as multi-agent systems in which a disrupted component becomes an irrational agent, either passive (an accidental fault) or active (adversarial, such as a cyber-attack).
Theodore Rappaport
New York University, NY·Engineering
Design and prototype a wideband liquid crystal-based reconfigurable intelligent surface (RIS) at 142 GHz, first passive and then active, using amplifier-integrated substrate-integrated waveguide elements.
Junshan Zhang
University of California-Davis, CA·Computer Science
Build CarDreamer, an open-source ecosystem of modular tools for world-model-based self-supervised learning in autonomous driving, through community scoping and planning with academic and industry users.
The school
University of Nevada-Reno: 11 labs on PI Pages, $194M a year on research, 186 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.