Efficient organization of the emergency evacuation of the population under the threat of catastrophic situations (like extreme flooding) in a limited amount of time requires the use of a predictive modeling and special control techniques in near real-time.
We know how to simulate the process of evacuating the populations of large urban areas in emergency situations, changing the possible scenarios for people’s movements. available ways and the scenario of people movements. For this purpose, we use the a model of traffic flow on a dynamic graph of the road network, the availability of which is associated with the crisis phenomena (for example, the progress of inundation during the flood).
This model allows one to study different scenarios, including the use of personal alerting methods, and provide decision support for the organization responsible for the evacuation process in a a critical situations. As an illustration, we compare four evaculation scenarios of the evacuation of people using personal vehicles during a catastrophic flood in St. Petersburg, which whose levels on average, is are met or exceeded once in a thousand years, in the case of a hypothetical failure of Flood Prevention Facility Complex. We consider various options of the behavior of individuals from the trivial, for example, following totaking the nearest bridge (top left), to "intelligent" routing, that takesing into account the current traffic and flood propagation (bottom right).
Distributed simulation of city inundation by coupled surface and subsurface porous flow for urban flood decision support system // Procedia Computer Science. — 2013. — Vol. 18. — pp. 1046-1056.
Towards a framework for simulation-based evaluation of personal decision support systems for flood evacuation // 14th International Multidisciplinary Scientific Geoconference SGEM 2014. GeoConference on Informatics, Geoinformatics and Remote Sensing.Conference Proceedings. — 2014. — Vol. 1., Issue 2. — pp. 883-892.