Embedded Files
My research interests lie broadly in control and optimization, incentive design, large-scale systems, and risk-averse/ safe learning. Specific topics that I have been dedicated to include:
♢ Incentive design and games,
♢ Control and optimization on manifolds/networks,
♢ Moment-based modeling and control of large-scale systems,
♢ Robust and active learning.
Incentive design for strategic agents
Incentive design for strategic agents
Modern infrastructure often interacts with strategic agents. Electricity consumers decide when to use power, drivers choose routes, and data owners determine what information to contribute. In each case, a central planner cannot directly prescribe the agents’ actions but instead influences them through incentive signals such as pricing, subsidies, tolls, and taxes. Incentive design studies how such signals should be chosen to align individual strategic behavior with system-level objectives.
The research studies incentive design as a feedback loop. Planners observe strategic agent responses, learn about their preferences, and update incentive mechanisms to steer the Nash equilibrium toward a desired collective objective, such as performance, welfare, fairness, or safety. The goal is not only to compute incentives when the model is known, but to adapt incentives online under information asymmetry and provide performance guarantees. [2025], [2026], [2026].
The work creates a natural connection between control, optimization, online learning, game theory, and multi-agent systems.
Control on manifolds & intrinsic formation
Control on manifolds & intrinsic formation
Control and optimization on manifolds in a distributed manner have become timely issues within the context of large-scale systems. It is because by identifying and exploiting the underlying geometric structures in high-dimensional data and states, one can dramatically reduce the complexity of the large-scale problems, and therefore make distributed algorithms more informationally efficient and computationally tractable.
We employed this line of work in differentiable manifolds and particularly in the n-sphere S^n and Lie group SO(3), which are entailed in numerous rigid-body attitude applications. Based on our study of the distributed algorithms of state consensus (i.e., distributed minimization of deviations between individuals) in non-Euclidean spaces, we further investigated formation problems for multiple rigid bodies.
Unlike most studies in the formation control literature, the new formation mechanism we proposed does not require any shape reference signal pre-given in the control protocol. Instead, the desired formation patterns are constructed totally based on the geometric properties of the configuration space and the designed connection topology. This type of formation is referred to as Intrinsic Formation Control. Moreover, the research of distributed coordination on manifolds is of not only theoretical but also practical significance in many applications, such as satellite constellation maneuver, drone cluster coordination, and multi-robot systems control. [Arxiv], [Arxiv], [2017], [2018], [2018].
Modeling & control of large-scale systems by moments
Modeling & control of large-scale systems by moments
Large-scale networked systems typically consist of a large number of connected agents, often too large for modeling each agent individually. Moreover, in many cases, the agents are exchangeable, and distinguishing each individual may not even be desirable. Therefore, we proposed a novel approach to characterizing the evolution of a network by a sequence of moments. In general, moments represent statistics and also represent realistically measured quantities of a cohort. By properly selecting kernel functions, the corresponding moments carry enough macro-scale information, such that the collective behavior of the networked system can be accordingly reconstructed.
The method provides a tractable approach to building the dynamics for a general moment sequence. In addition, the resulting moment systems typically permit less intricate and lower-dimensional dynamics, which considerably reduces computational complexity for controlling and analyzing large-scale systems. The method applies to a wide range of applications, such as multi-agent systems, crowd dynamics, opinion dynamics, and many problems pertinent to collective behavior occurring in biology, economics, and social sciences. [Arxiv]
Control a pedestrain crowd by the moment method
Network coordination with privacy
Network coordination with privacy
In many Internet of Things applications, privacy is a hotly debated topic and subject to several regulations in recent years, e.g., GDPR, PIPEDA, and CCPA. When sensitive data is exchanged between parties, there is no guarantee that it remains private or that it isn't used for nefarious purposes. We study privacy-preserving approaches for network coordination, by which all nodes in a network can achieve the collective goals, such as consensus, collaboratively optimizing a function, and training in federated learning, but without exposing the individual privacy to other parties. In contrast to methods that use cryptography, we pursue the preservation of participants' privacy in information exchanges over networks from the perspective of dynamical systems theory, which offers greater computational efficiency and application flexibility. [Arxiv], [Arxiv]
Google Sites
Report abuse