The Effect of Information on the Performance of a Multiagent System for Post-Disaster Evacuation

The paper presents a new approach of simulating various post-seismic evacuation scenarios for people in Iasi city, Romania, by using a multiagent system. The emergency scenarios consider that a key point of Iasi road network may become non-functional because of structural damage, and may interrupt the evacuation process. Thus, the population groups modelled by agents are impelled to look for alternative routes. A comparison between the prior-and post-seismic event situations is presented. The focus is on the effect of the amount of information the agents have about their destinations, i.e. shelters or hospitals: the way in which the evacuation time decreases as more information becomes available. Some emergent, and sometimes counter-intuitive outcomes are discussed. The results can be considered to serve as an evaluation scheme of the post-emergency situation for the considered locations. The use of agents facilitates the finding of appropriate solutions that reduce the population's evacuation time.