Modelling of crowd evacuation behavior considering the effects of drunken gait

The presence of drunk crowds complicates the evacuation process compared to sober crowds. We construct a Floor-Field Cellular Automata (FFCA) model considering the role of drunken pedestrian perturbations to investigate the impact of drunken gait on pedestrian dynamics. A random floor field is constructed to simulate the drunken gait of pedestrians, considering the role of surrounding pedestrians and the random movement behaviour of "drunken walk". A drunkenness factor is proposed to modulate the randomness of drunken gait, and a velocity model is formulated to account for the intermittent speed fluctuations observed in drunken pedestrians. In addition, a probabilistic model incorporating the coupled effects of drunkenness level and movement speed is proposed to address the pedestrian conflict. The FFCA model can reproduce some real features of the drunken evacuation process to some extent. For example, (1) Drunken pedestrian staggers while moving. (2) Blockage occurs at the gathering place of drunk pedestrians. (3) Drunken crowds are pushed to move by the flow of the crowd. The results show that drunken gait disrupts the orderly movement of pedestrians, whereas the presence of sober pedestrians can foster the formation of adaptive orderly behaviors, maintaining a orderly movement structure. When drunken pedestrians are located at the rear of the crowd flow, evacuation efficiency is higher in the early stage but the overall efficiency of the entire evacuation process is the lowest. Conversely, when distributed in the front and middle, although the gait of drunken people hinders in the early stage evacuation, sober pedestrians in the rear can act as guides and improve the overall evacuation efficiency.