Simulation of migration paths using agent-based modeling: The case of Syrian refugees

The Syrian civil war, which has been ongoing for over a decade, has caused a massive migration wave in the Middle East. Turkey has been the most preferred destination for Syrian refugees seeking security. The primary purpose of our study is to develop a migration tool to simulate the paths of Syrian refugees en route to Turkey. In this way, we aim to facilitate humanitarian aid organizations to deliver services to refugees in future migration instances. We limited the scope of the study to a crisis moment. We simulated the southeastern Idlib aerial bombardment on December 16, 2019, and observed the outcomes of the sudden conflict-driven migration movements of refugees who started to migrate in a one-month period from December 16, 2019 to January 15, 2020. Due to the lack of reliable data on displacement during the crisis, we restricted the destination options to official refugee camps in Turkey. We developed an agent-based model (ABM) based on two attributes of refugee groups: risk sensitivity and agent type, which include four different information types. We expanded the A* algorithm with a cost metric to calculate the weighted average of distance and risk to a destination point. The proposed ABM has the flexibility to be easily adapted to migration instances in another geography and to enable various analyses under different parameters. The case study examines the southern Idlib crisis with five different scenarios. We obtained the total number of refugees, migration paths, camp occupancy rates, and heat maps of the densely populated regions for each scenario. We compared the obtained outputs with the situation reports and the official statements of the period and acquired several meaningful insights for policymakers.