Andrew M. Fraser
XCP-8, Los Alamos National Laboratory

Stephen A. Andrews
XCP-8, Los Alamos National Laboratory

Abstract

Many physical processes are modeled by unspecified functions. Here, we introduce the F\_UNCLE project which uses the Python ecosystem of scientific software to develop and explore techniques for estimating such unknown functions and our uncertainty about them. The work provides ideas for quantifying uncertainty about functions given the constraints of both laws governing the function's behavior and experimental data. We present an analysis of pressure as a function of volume for the gases produced by detonating an imaginary explosive, estimating a best pressure function and using estimates of Fisher information to quantify how well a collection of experiments constrains uncertainty about the function. A need to model particular physical processes has driven our work on the project, and we conclude with a plot from such a process.

Keywords

python, uncertainty quantification, Bayesian inference, convex optimization, reproducible research, function estimation, equation of state, inverse problems

DOI

10.25080/Majora-629e541a-001

Bibtex entry

Full text PDF

Copyright © 2016 Andrew M. Fraser et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.