Agent-Based Rescue Simulation During Airstrikes with an Emphasis on Task Allocation Between Groups

Multi-agent systems (MASs) provide the possibility of prediction in critical situations by simulating the components of a complex system as intelligent agents and implementing its decision-making parameters. The main objective of this paper is to simulate the search and rescue operations for airstrikes (simulation of search, rescue and medical agents) and to increase the effectiveness of rescue teams in critical situations. Implementation of this system was carried out in district 3 of Tehran in three main phases: 1) prioritization of the area with the technique for the order of preference by similarity to the ideal solution (TOPSIS) in the GIS environment, 2) simulating the operating environment and establishing cooperation between the rescue groups by the AnyLogic software, and 3) creating scenarios and analyzing the different results. The infrastructure priorities’ map showed that 21 percent of the city’s blocks in the region were priority number one to enemy air attack and the other 18, 29, 22 and 10 percentages were in the second to fifth priorities, respectively.  Implementation of a simulator system revealed that, by doubling the number of rescue forces in different scenarios, the average time of operation was decreased by 47% and the number of fatalities decreased by 9%. The results indicate the value of using multi-agent systems in designing a search and rescue simulator system and modeling the decision-making components. The output of this article can be used to manage, decide, and predict the extent of the damage caused by air attacks.