Bourne, Jonathan;
(2021)
Strain Elevation Tension Spring embedding and Cascading failures on the power-grid.
Doctoral thesis (Ph.D), UCL (University College London).
Preview |
Text
SETSE_Thesis_Library.pdf - Accepted Version Download (11MB) | Preview |
Abstract
Understanding the dynamics and properties of networks is of great importance in our highly connected data-driven society. When the networks relate to infrastructure, such understanding can have a substantial impact on public welfare. As such, there is a need for algorithms that can provide insights into the observable and latent properties of these structures. This thesis presents a novel embedding algorithm: the Strain Elevation Tension Spring embedding (SETSe), as a method of understanding complex networks. The algorithm is a deterministic physics model that incorporates both node and edge features into the final embedding. SETSe distinguishes itself from most embeddings methods by not having a loss function in the conventional sense and by not trying to place similar nodes close together. Instead, SETSe acts as a smoothing function for node features across the network topology. This approach produces embeddings that are intuitive and interpretable. In this thesis, I demonstrate how SETSe outperforms alternative embedding methods on node level and graph level tasks using networks made from stochastic block models and social networks with over 40,000 nodes and over 1 million edges. I also highlight a weakness of traditional methods to analysing cascading failures on power grids and demonstrate that SETSe is not susceptible to such issues. I then show how SETSe can be used as a measure of robustness in addition to providing a means to create interpretable maps in the geographical space given its smoothing embedding method. The framework has been made widely available through two open source R packages contributions, 1) the implementation of SETSe ("rsetse" on CRAN), and 2) a package for analysing cascading failures on power grids.
| Type: | Thesis (Doctoral) |
|---|---|
| Qualification: | Ph.D |
| Title: | Strain Elevation Tension Spring embedding and Cascading failures on the power-grid |
| Event: | UCL (University College London) |
| Open access status: | An open access version is available from UCL Discovery |
| Language: | English |
| Additional information: | Copyright © The Author 2021. Original content in this thesis is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0) Licence (https://creativecommons.org/licenses/by-nc/4.0/). Any third-party copyright material present remains the property of its respective owner(s) and is licensed under its existing terms. Access may initially be restricted at the author’s request. |
| UCL classification: | UCL UCL > Provost and Vice Provost Offices > UCL BEAMS UCL > Provost and Vice Provost Offices > UCL BEAMS > Faculty of the Built Environment UCL > Provost and Vice Provost Offices > UCL BEAMS > Faculty of the Built Environment > Centre for Advanced Spatial Analysis |
| URI: | https://discovery.ucl.ac.uk/id/eprint/10134101 |
Archive Staff Only
 |
View Item |
