Information-Geometric Approach For Data-Driven Multiscale Simulations

Abstract

The goal of this project is to construct accurate and efficient hybrids-multiscale algorithms while addressing three inter-connected challenges: how to construct a coarse-scale model that is “consistent” (in a probabilistic ) with its fine-scale counterparts, how to propagate noise and stochastic fluctuations inherent in fine-scale simulations into the coarse-scale simulations, and how to assimilate both fine-scale (simulated and/or experimental) data into the coarse-scale model and coarse-scale data into the fine-scale model.

Document Details

Document Type
DoD Grant Award
Publication Date
Sep 19, 2018
Source ID
FA95501810474

Entities

People

  • Daniel M. Tartakovsky

Organizations

  • Air Force Office of Scientific Research
  • Stanford University
  • United States Air Force

Tags

Readers

  • Computational Modeling and Simulation
  • Neural Network Machine Learning.
  • Ocean-Atmosphere Mesoscale Modeling, Data Assimilation, and Flux Boundary Layers