High Performance Computational Structure for Parameter Estimations of Parametric Partial Differential Equations

Abstract

One of the great scientific challenges of this century is to accurately describe complex processes such as climatechange, contamina""nt flow, genomics, and even social media and finance. The main obstacle to our understanding of these processes is that they depend" on many parameters or variables. This not only complicates their mathematical description but inhibits the use of modern computatio"nal tools for their accurate prediction. New mathematical ideas, based on sparsity and model reduction, have been put forward by the" PIs and other researchers to determine theimportance of the various parameters and to derive simpler models that accurately describe the underlying process. This in turn leads to more accurate and less costly computational models that can be implemented within today~s computational resources. Yet validation of less costly computational models or relevant physical applications require paralle"l implementations using hundreds of thousands of independent singlecore computations, which greatly exceeds the resources currently" available to the Principal Investigator.

Document Details

Document Type
DoD Grant Award
Publication Date
Sep 29, 2017
Source ID
N000141712908

Entities

People

  • Ronald DeVore

Organizations

  • Office of Naval Research
  • Texas A&M University
  • United States Navy

Tags

Readers

  • Computational Modeling and Simulation
  • Distributed Systems and Data Platform Development