Application of cloud computing and big data in three-stage dynamic modeling of disaster relief logistics and wounded transportation: a case study.

Abstract

Collecting and sharing information about affected areas is an important activity for optimal decision-making in relief processes. Defects such as over-sending some items to affected areas and mistakes in transferring injured people to medical centers in accidents are due to improper management of this information. Because cloud computing as a processing and storage platform for big data is independent of the device and location and can also perform high-speed processing, its use in disasters has been highly regarded by researchers. In this environment, a three-stage dynamic procedure for evacuation operations and logistics issues is presented. The first stage of the proposed model is image processing and tweet mining in a cloud center in order to determine the disaster parameters. In stage II, a mixed-integer multi-commodity model is presented for the relief commodity delivery, wounded people transportation with capacity constraints, and locating of the possible on-site clinics and local distribution centers near disaster areas. In stage III, by using a system of equations, detailed vehicle load/unload instructions are obtained. Finally, the effectiveness of the proposed model on the data of an earthquake disaster in Iran is investigated. The results of comparing the proposed approach with a two-stage algorithm show that the total number of unsatisfied demand for all types of commodities in the proposed approach was better than the other. Also, the number of survivors in the three-stage model is significantly higher than in the two-stage one. The better performance of the proposed algorithm is due to the fact that online data is continuously available and that decisions such as sending relief items and dispatching are made more effectively.