Multi-regime modelling of large congregations

Abstract

Crowd gatherings at religious occasions, fairs, transport terminals and so on can create severe threats for the crowd due to the high density of people in a certain space, with adverse outcomes such as crowd stampedes. Analyses of stream flow parameters describe the relationships among the characteristics of crowd traffic streams. A regime model chosen based on crowd density has an impact on the predicted crowd characteristics at both the micro-level (individual behaviours in a crowd) and macro-level (overall crowd behaviour). Two-regime and three-regime models were used in this study to characterise crowd traffic in different regimes such as uncongested flow, transitional flow and congested flow. A model was developed based on the better fitness of predicted data to actual observed data. Data from the Krishna Pushkaralu festival were used for the model development and the developed model was validated using data from the Medaram festival. The three-regime model was found to provide a better fit than the single- and two-regime models and it was therefore concluded that the three-regime model can be used to analyse crowd behaviour and effectively manage crowds at any large congregation.