Software

This file contains the routines necessary to generate networked surrogate models for multifidelity information fusion. It forms the basis of the routines used to generate the results of this paper

• Gorodetsky, A.A., Jakeman, J.D. & Geraci, G. MFNets: data efficient all-at-once learning of multifidelity surrogates as directed networks of information sources. Comput Mech (2021). https://doi.org/10.1007/s00466-021-02042-0, http://arxiv.org/abs/2008.02672

The code can be found on GitHub. Autogenerated documentation is available here.

The Compressed Continuous Computation (C3) library (in C) enables computation with multivariate functions. Many multilinear algebra computations are included. Common tasks include addition, multiplication, integration, differentiation, and approximation of multivariate functions.

Doxygen based documentation is available here

The C3SC add-on to the C3 package provides utilities for solving general stochastic optimal control problems. The included algorithms can handle non-affine controls and non-quadratic costs. The underlying theory is based on low-rank representations of value functions.

GPEXP is a software package, written in python2.7, for performing experimental design in the context of Gaussian process (GP) regression. Experimental design may be performed for a variety of cost function specifications. Currently supported cost functions include those based on integrated variance, conditional entropy, and mutual information. GPEXP may also be used for general purpose GP regression. Currently supported kernels include the isotropic and anisotropic squared exponential kernel, the isotropic Matern kernel, and the Mehler kernel. Additional kernels may be easily specified. GPEXP also includes optimization routines for estimating kernel hyperparameters from data.

The Dakota project is the industry state-of-the-art software package for complex analysis having to do with uncertainty quantification of computational models. The recent releases have included C3 as a third party library. As a result, it is not possible to perform extensive UQ analysis with functional tensor trains.

Many of the multifidelity uncertainty quantification algorithms that we have published are implemented in PyApprox. In particular, this package provides implementation of ACV and MFNets. The package also includes tools for wide ranging areas of uncertainty quantification

This package implements a very wide range of multi-fidelity (multi-model) Monte Carlo sampling algorithms. In particular it provides extensive realizations as well as extenstions of the Approximate Control Variate. The GitHub is here.