https://github.com/adrianlepp/nonlinear-lodegp-control

https://github.com/adrianlepp/nonlinear-lodegp-control

Science Score: 36.0%

This score indicates how likely this project is to be science-related based on various indicators:

  • CITATION.cff file
  • codemeta.json file
    Found codemeta.json file
  • .zenodo.json file
    Found .zenodo.json file
  • DOI references
  • Academic publication links
    Links to: arxiv.org
  • Academic email domains
  • Institutional organization owner
  • JOSS paper metadata
  • Scientific vocabulary similarity
    Low similarity (10.4%) to scientific vocabulary
Last synced: 10 months ago · JSON representation

Repository

Basic Info
  • Host: GitHub
  • Owner: adrianLepp
  • License: mit
  • Language: Python
  • Default Branch: main
  • Size: 86.9 KB
Statistics
  • Stars: 0
  • Watchers: 0
  • Forks: 0
  • Open Issues: 0
  • Releases: 0
Created about 1 year ago · Last pushed about 1 year ago
Metadata Files
Readme License

README.md

nonlinearlodegpcontrol

This project is based on the LODE-GP and was developed in my master thesis Physics-informed Gaussian Process Regression: Applications in Modelling and Control of Nonlinear Systems.

For the part about Model Predictive Control (MPC), check the pulbication Physics-informed Gaussian Processes for Model Predictive Control of Nonlinear Systems. Everything else is not publicated yet.

Check the examples, on how to use the implemented methods.

Requirements

For installation run

bash git clone git@github.com:adrianLepp/nonlinear-lodegp-control.git pip install -e ./nonlinear-lodegp-control

SageMath (https://www.sagemath.org/) needs to be installed seperately:

```bash

Example install via conda (recommended):

conda install -c conda-forge ```

Remark on gpytorch

The heteroskedastik noise for a multi-dimensional GP (as it is used for MPC), is not implemented in gpytorch yet. See https://github.com/cornellius-gp/gpytorch/issues/901 for proposed implementations. One possible implementation is present in this repository.

Calling the likelihood on another distribution than the training data is also not possible, due to a bug that is not fixed yet. See https://github.com/cornellius-gp/gpytorch/issues/2630 for a description of this bug and a patch that can be applied to the source code. However, it is normally not necessary to call the likelihood on something else than the training data.

Owner

  • Login: adrianLepp
  • Kind: user

GitHub Events

Total
  • Push event: 4
  • Create event: 4
Last Year
  • Push event: 4
  • Create event: 4

Dependencies

pyproject.toml pypi
  • einops *
  • gpytorch *
  • matplotlib *
  • numpy *
  • sage *
  • scipy *