CISE Research Instrumentation: A High-Performance Computing Facility for Experimental Algorithms

点击翻译按钮获取中文摘要

9720676 Baruah, Sanjoy K. University of Vermont and State Agricultural College CISE Research Instrumentation: A High-performance Computing Facility for Experimental Algorithmics This research instrumentation grant enables research in several projects: - Computing Euclidean Steiner Minimum Trees,- Research on Blocking Behavior of Multicast Networks,- Feasibility Analysis via Computation-Intensive Simulations, - Finite Sample Analyses of Nearest Neighbor Algorithms, and - Reliability of Large Scale Disk Arrays. To support the aforementioned projects, the newly-formed Department of Computer Science at the University of Vermont will purchase high-performance computing equipment. The equipment will be used in particular for: - Developing efficient algorithms for generating the shortest network interconnecting a given set of points.- Investigating the blocking behavior of multicast networks.- Developing simulation algorithms for the feasibility analysis of hard-real-time application systems.- Developing and evaluating nearest-neighbor algorithms for pattern recognition.- Exploring techniques for enhancing the reliability of disk arrays. Each project will explore the impact of the recent dramatic advances in computing technology upon algorithm design issues in a particular application domain. New techniques, methodologies, and algorithms will be designed which make extensive use of large-scale simulation and/ or combinatorial search. The equipment made available through this grant will be used to both perform the computational experiments necessary to guide the research, and to validate the findings thus obtained.