TRANSIMS: Applications and Development Workshop
Simulation of building evacuation exit rates for TRANSIMS
Reinaldo Moraga
Department of Industrial and Systems Engineering
Northern Illinois University
590 Garden Road
DeKalb, IL 60115-2854
(815) 753-1442
Abstract Biography
List of Authors
================
Reinaldo Moraga
Department of Industrial and Systems Engineering
Northern Illinois University
=========
This presentation deals with the topic of building evacuation modeling for TRANSIMS large scale simulation models in major metropolitan cities. The idea consisted of creating a generator of building exit times in order to fine tune the current version of TRANSIMS trip tables. A major challenge in order to solve this problem is the nonexistent data from past experiences of real world building evacuations. Literature does not provide this type of specific information. Therefore, it was necessary to create an agent-based model, using SIMULEX, to simulate the evacuation of a building and evaluate alternative escape modes to obtain data. A drawback on this modeling tool is that it highly depends on building layout and as the number of floors increases modeling efforts becomes really time-consuming. Moreover, this limitation impedes to easily interface SIMULEX with TRANSIMS. For this reason, a Monte-Carlo simulation algorithm was designed using the data collected from the agent-based building evacuation model. This algorithm acts like a meta-model of SIMULEX software, where some input variables (e.g., number of floors, type of people inside of the building) are entered and it provides exit rates in order to calculate inter-exit times. The algorithm can be easily implemented and called by TRANSIMS. It is based a Non-stationary Poisson process since arrival rates to the building exit points are not constant but changing over time. The Monte-Carlo simulation algorithm was tested for 10, 15, 25 floors and was capable to replicate similar exit rates as SIMULEX does. In addition, the algorithm was used to predict exit rates for a 100-floors building. In this presentation, the design of this Monte-Carlo simulation algorithm will be presented; validation issues and future research ideas will be discussed.
==========
Reinaldo Moraga, Ph.D., is an Assistant Professor in the Department of Industrial and Systems Engineering at the Northern Illinois University. He received a B.Sc. degree in Industrial Engineering in 1993 from Universidad del Bio, in Chile, and has degrees of M.Sc. and Ph.D. in Industrial Engineering from University of Central Florida. He has written more than 30 articles in technical proceedings and journals. His research interests are Simulation, Operations Research, Metaheuristics, and Disaster Operations Management.