IEEE Transactions on Emerging Topics in Computational Intelligence | 2021

Transport Management for Evacuation of Victims

 
 

Abstract


The metropolitan area experiences disaster due to large population density, urbanization and heterogeneous demands at different time periods and places, like accidents, sudden vehicle breakdowns, congestions, lane blockages, earthquakes, flood and hurricanes. These disasters cause severe property damages, economic losses and losses of life. Therefore, in order to reduce the effects of disasters an effective evacuation management of victims from disaster zone is necessary. The evacuation during and after the disaster is difficult due to several unforeseen hurdles. We need a dynamic and adaptive transport management scheme to suite the requirements. In this paper, we propose Emergent Intelligence (EI) technique based transport management for evacuation of victims from disaster zone to safe zone. The EI technique makes static agent (resided at depot in a disaster zone) to logically divide disaster zones into multiple zones and sub-zones. Static agent creates and dispatches mobile agents to these zones and sub-zones, and EI technique makes group communication among mobile agents and static agent, during group communication they collect, analyze and share estimated victims density, resources availability, resources allocated, relevant routes availability, and boundary of disaster. The static agent at a depot estimates guaranteed routes, evacuation exit points, fraction of victims arrive at each exit points, average travel time to exit evacuees from disaster location to exit points. A case study is presented to test the proposed system transport allocation to different routes depending upon population of victims for evacuation management in a disaster zone. The proposed system reduces average delay propagation (i.e., 4 sec.) of disaster information compared to the traditional system (i.e., 30 sec.), minimizes the average evacuation time (i.e., at max. 10 minutes) required for evacuating victims from disaster zone to safe zone, and also uniformly distributes victim among evacuation exit points.

Volume 5
Pages 426-441
DOI 10.1109/TETCI.2019.2940832
Language English
Journal IEEE Transactions on Emerging Topics in Computational Intelligence

Full Text