Regresar

Estimation of evacuation time in an emergency in an educational center based on a neural network approach

Abstract:

The abstract should not exceed 150 words. You can use this document as the basis for final editing of the article to submit. The estimation of emergency evacuation time in public places becomes a relevant aspect for disaster management. This research focuses on the correct estimation of emergency evacuation response rates considering macroscopic elements, behavior assumptions, circulation allocation, supply and demand management strategies in human performance and structural conditions of the sites. For the pbkp_rediction of evacuation time, an Artificial Neural Network with a Multilayer Perceptron called Back Propagation was used. The feeding data for the Artificial Neural Network were taken from 10 practical evacuation exercises carried out under maximum working conditions to a population of 18,000 occupants of four university campuses of the Technical University of Ambato. After the application of the algorithm in this case study, values over 80% certainty of the real values were checked.