A SURVEY ON TOWARD EFFECTIVE RESPONSE TO NATURAL DISASTERS: A DATA SCIENCE APPROACH

Abstract

Natural catastrophes have the potential to destroy large portions of infrastructure and kill thousands of people. Both the populace and the government find it challenging to deal with these situations. Particular attention must be given to the following two difficult problems: find a workable solution first evacuating people, then rebuilding homes and other infrastructure. Then, a successful recovery plan that prioritises the reconstruction of damaged areas and the evacuation of people can be a game-changer for overcoming those horrible circumstances. In this light, we introduce DiReCT, a method based on I a dynamic optimization model created to quickly develop an evacuation plan of an earthquake-stricken area, and ii) a double deep Q network-based decision support system capable of effectively guiding the rebuilding of the affected areas. The latter operates by taking into account the needs of the many stakeholders (such as citizens' social benefits and political priorities) as well as the resources available. The foundation for both of the aforementioned solutions is a specialized geographic data extraction Method called "GisToGraph," which was created expressly for this use. We used extensive GIS data, information on the vulnerability of urban land structures, and the historical city centre of L'Aquila (Italy) to test the applicability of the entire strategy.