Publication detail

Real-time RSET prediction across three types of geometries and simulation training dataset: A comparative study of machine learning models

UHLÍK, O. OKŘINOVÁ, P. TOKAREVSKIKH, A. APELTAUER, T. APELTAUER, J.

Original Title

Real-time RSET prediction across three types of geometries and simulation training dataset: A comparative study of machine learning models

Type

journal article in Web of Science

Language

English

Original Abstract

Agent-based evacuation models provide useful data of the evacuation process, but they are not primarily designed for use during an emergency. The paper aims to test predicting RSET using a surrogate ML model trained on a simulation dataset with 60 samples. A total of 9 machine learning algorithms were tested on 3 simple geometries: bottleneck, stairway and walkway. A set of 7 spatial features was used to train the surrogate models. The results showed a relatively good ability of Artificial Neural Network to learn in scenarios involving bottlenecks and stairways, with an R2: 0.99 on the testing dataset. In the walkway scenario, all models experienced a significant drop in performance, with Gradient Boost performing the best (R2: 0.92). The paper demonstrated ability to generalize effectively in bottleneck-type tasks with training on a relatively small dataset containing spatial parameters obtainable in real-time from camera systems.

Keywords

Evacuation modeling, Machine learning, Required safe egress time, Agent-based models

Authors

UHLÍK, O.; OKŘINOVÁ, P.; TOKAREVSKIKH, A.; APELTAUER, T.; APELTAUER, J.

Released

24. 5. 2024

Publisher

Elsevier

Location

Netherlands

ISBN

2666-1659

Periodical

Developments in the Built Environment

Year of study

18

Number

100461

State

United Kingdom of Great Britain and Northern Ireland

Pages count

13

URL

https://www.sciencedirect.com/science/article/pii/S266616592400142X?via%3Dihub

Full text in the Digital Library

http://hdl.handle.net/11012/249353

BibTex

@article{BUT188671,
  author="Ondřej {Uhlík} and Petra {Okřinová} and Artem {Tokarevskikh} and Tomáš {Apeltauer} and Jiří {Apeltauer}",
  title="Real-time RSET prediction across three types of geometries and simulation training dataset: A comparative study of machine learning models",
  journal="Developments in the Built Environment",
  year="2024",
  volume="18",
  number="100461",
  pages="13",
  doi="10.1016/j.dibe.2024.100461",
  issn="2666-1659",
  url="https://www.sciencedirect.com/science/article/pii/S266616592400142X?via%3Dihub"
}