Resolution of deadlocks in a fine discrete floor field cellular automata model—modeling of turning and lateral movement at bottlenecks

Abstract

For a parallel scheme, finer discretization of space in cellular automaton models of pedestrian dynamics can lead to new problems, e.g. the formation of deadlocks. In order to deal with such problems in the floor field cellular automata (FFCA) model, the turning and lateral movement of occupants at a bottleneck are considered. Simulations of the proposed model are found to fit well with experiments. The turning breaks the deadlock and releases the congestion at the bottleneck. The deadlock frequency decreases and the free travel time increases with the turning probability at the bottleneck. For slow lateral movement, pedestrians take more time to reach the exit, and the peak flow is first reached at a later time. As the discretization becomes finer, the crowd shape around the bottleneck changes gradually from semi-circle to semi-ellipse. Pedestrians in the finer discrete model seem to prefer queuing in a line rather than making a detour. The increase of lateral velocity may lead to more frequent conflicts among pedestrians.