GIS Data Driven Probability Map Generation for Search and Rescue Using Agents

Abstract

Predicting the final resting location of a missing person is critical for search and rescue operations with limited resources. To improve the accuracy and speed of these predictions, simulated agents can be created to replicate the behavior of the missing person. In this paper, we introduce an agent-based model, to simulate various psychological profiles, that move over a physical landscape incorporating real-world data in their decision-making without relying on per-location training. The resultant probability density map of the missing person's location was the result of a combination of Monte Carlo simulations and mobility-time-based sampling. General trends in the data were comparable to historical data sets available. This work presents a flexible agent that can be employed by search and rescue that easily extends to various locations.