Evacuation Behaviour Prediction Through Space Syntax-Based Agent Simulation at Sentra Tahu Cibuntu Bandung

Abstract The development of the theory of spatial syntax is currently being employed in a number of different contexts, including the creation of evacuation plans. In order to predict the movement and utilisation of individuals within buildings and urban environments, space syntax derives a set of analytical measures pertaining to the configuration of streets. In this research, space syntax theory is employed as the foundation for simulating agents in the context of fire evacuation and wayfinding in segregated spaces in informal settlements that evolve spontaneously and incrementally. The research sample was the Sentra Tahu Cibuntu (Cibuntu Tofu Center). This is an informal settlement with an unplanned spatial configuration. The method employed is a space syntax-based agent simulation utilising DepthMapX 0.8 software in the most segregated area. The objective is to predict evacuation pathways in areas susceptible to natural disasters, crime, and health-related hazards. The findings demonstrate that residents consistently gravitate towards open areas. In this context, open spaces emerge as generators of human movement based on their visual perspective.