This repository contains the companion code repository for Quantifying Uncertainty in Subsurface Systems by Céline Scheidt, Lewis Li, and Jef Caers (John Wiley & Sons).
This repository implements the various UQ strategies discussed in the book. The source code for the algorithms can be found under the src folder. The codebase was developed in MATLAB
For illustrative purposes, a set of Jupyter tutorials have been prepared. They are as follows:
- Dimension Reduction: Showcase of various dimension reduction techniques discussed in Chapter 3.
- DGSA: Implementation of Distance Based Sensitivty Analysis from Chapter 4.
- Bayesian Evidential Learning. Methodology discussed in Chapter 7, implemented for the Libyan Oil Reservoir case.
- SIR: The Sequential Importance Resampling methodology from Chapter 7 applied to the same Libyan Oil Reservoir.
These tutorials can be viewed directly in the browser, or download and re-run.
To re-run the examples, the following dependencies must be met:
- MATLAB
- MATLAB Engine API for Python
- jupyter or can be installed as part of Anaconda package
- matlab_kernel for interfacing MATLAB with jupyter
Once installed, jupyter can be started from the command line with
jupyter notebook
Navigate to the tutorials folder and select appropiate tutorial to load.
This repository is released under the MIT License.