GPU-accelerated mathematical optimization with applications to big data, computer modeling, and simulation
Abstract
This proposal seeks funds to purchase a NVIDIA DGX-2 server for the purpose ofsolving large, interdisciplinary problems that arise in mathematical optimization, statistics, and engineering contexts. Such problems, which due to their complexity cannot be solved through standard software and/or hardware, are prevalent in network security and vulnerability, combinatorial optimization problems in scheduling, sequencing, assignment, and resource allocation, bigdata analysis for object classification, statistical imaging, oceanic modeling and simulation, and deep reinforcement learning for Unmanned Aerial Vehicles. To successfully attack such problems, a multidisciplinary research team with complementary expertise seeks to incorporate a DGX-2 server as a special node in Clemson~s Palmetto cluster. Two of the investigators are currentlyfunded on ONR grants, two others are funded on grants from NIH and the Simons Foundation, and a proposal is pending with NSF. These grants conduct fundamental research, but powerful high-performance computing is needed to effectively leverage the theoretical findings to solve the problems under consideration. The requested equipment will be instrumental in the success of two interdisciplinary research centers, and be a valued asset for general research and instructions. Realizing that problems confronting society encompass many traditional academic disciplines, Clemson University has placed emphasis on interdisciplinary research. A consequence of this emphasis is the recent formation of two centers: the Statistics and Mathematics Consulting Center (SMCC) and the Clemson Operations Research Institute (CORI). The SMCC provides a portal to the community that facilitates collaboration and consultation on data analysis, statistical problem solving, scientific computing, and mathematical modeling. The CORI is a center for research relative to mathematical optimization. Every investigator on this proposal is affiliated with at least one of these institutes, and the requested equipment will form the computational foundation for both centers. To fully utilize this computing platform to further science, priority access will be granted to the members that are actively working on defense-oriented research, and secondary access will be granted to the entire academic community of Clemson for both research and educational opportunities.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Apr 25, 2019
- Source ID
- N000141912295
Entities
People
- Yuyuan Ouyang
Organizations
- Clemson University
- Office of Naval Research
- United States Navy