Assessment of velocity variation algorithms in evacuation models with heterogeneously distributed factors

This study investigates the impact of velocity variation algorithms on evacuation simulations using two movement models: the Floor Field Cellular Automata (FFCA) model and the Natural Step Length Floor Field Cellular Automata (NSFFCA) model. The focus lies in understanding how different approaches to velocity variation affect pedestrian movement during evacuations. Traditionally, FFCA models reduce overall pedestrian speed by lowering movement probability at each time step. In contrast, the NSFFCA model integrates a natural step length based on individual speed, allowing pedestrians to adjust their pace by decreasing step size. The study examines model outputs under two sets of environmental factors: simulated fire-generated products and local pedestrian density. Both models incorporate the same empirical relationships to adjust pedestrian movement based on these factors, but differences in velocity variation algorithms influence simulation outcomes. It is expected that in evacuation scenarios with heterogeneously distributed elements, the differing speed variation approaches between models can influence pedestrian exposure to these factors and affect overall evacuation outcomes. In individual evacuation scenarios with heterogeneously distributed fire products, the FFCA model demonstrates greater run-to-run variation in evacuation time due to its velocity variation algorithm. Pedestrians encountering fire products reduced speed by decreasing movement probability per time step, potentially resulting in halted movement. This increased variability in pedestrian movement leads to prolonged exposure to fire products, impacting overall simulation outcomes. The NSFFCA model maintains a more consistent pace by adjusting step lengths, resulting in reduced variability. In congested exit scenarios, both models converge toward similar movement distances due to limited space, minimising differences in velocity variation.