Dr. Changqing Cheng
Institution: Binghamton University
Department: Systems Science Industrial Engineering
Proposed research ideas:
I have been focusing my research on data-driven modeling of various complex systems for monitoring and control purposes. I am interested in exploring the following ideas collaboratively with research teams/groups at DOE labs. (1) Data infusion: heterogeneous data sources are readily available, for example, image and time-domain data, structured and unstructured data, data sampled at different frequency, and data from sensor measurements and physical models. An effective way to fuse them seamlessly is needed to for anomaly detection, dynamic evolution tracking and process prognostics, in applications from energy production system to climate science. (2) Uncertainty quantification: As one of the ongoing research in my research team, a sparse representation-based polynomial chaos is used for the uncertainty quantification in the modeling of systems with multiple heterogeneous data sources (e.g., physical simulation model and sensing data). We are investigating the reliability in the tidal energy system. (3) Coarse-grained representation of spatiotemporal dynamics in complex networks: Complex networks can be used to represent spatiotemporal dynamics in the underlying systems, such as nano-synthesis process (atoms are the nodes, and bonds are the edges) and wind farm energy output. As the number of nodes increases, it is becoming a challenge to study those systems. A coarse-grained (CG) representation for large-scale complex networks is currently being investigated in my research team, which condenses a cluster of nodes into a CG site, and the interaction between the CG sites is derived according to the nodal interactions. This approach has the potential to scale up the analysis of real-world network problems, without sacrificing too much accuracy.
(1) Interdisciplinary research: The data-driven complex system modeling is more about interdisciplinary research. I have been involved in the study of complex systems from nanomanufacturing to healthcare to energy systems. This program offers a precious opportunity to expose myself to the collaborative leading-edge interdisciplinary research with computing scientists in DOE labs. Therefrom, I can build my momentum towards high-impact research. I regard this as a new starting pointing for my academic career. (2) It is also a good opportunity to network with other top-tier researchers and scientists, and other faculty peers in the community of computational science. This will further provide chance to explore potential research collaborations and sponsored projects. This network, connections and friendship will continue to benefit me throughout my career. (3) At the end of this program, I anticipate several high-quality publications, and most importantly, a well-organized collaborative research proposal, which will have the potential to address some of the most challenging issues in complex system modeling and control. This will be my research focus for the next 3 years.