dynamic network loading model for p

dynamic network loading model for pedestrian flows has been developed. It represents a macroscopic approach for
studying large-scale problems of crowd dynamics in presence of congestion, and is based on an isotropic approximation
to the continuum theory of pedestrian flows (Hughes, 2002). An analysis of various basic flow patterns, and the application
of the framework to two real case studies, have revealed a realistic behavior of the model under various conditions. In the
future, the developed framework may be improved and extended in several ways. First, the use of a stochastic fundamental
diagram may be considered (Nikolic´ et al., 2014). Such an approach would likely allow to explicitly model random fluctuations
in walking times as observed in the pedestrian tracking data presented in this work. Second, a multi-class framework
would allow considering different density-speed relations for different types of pedestrians such as ‘business travelers’,
‘commuters’, ‘seniors’, and so on (Cooper, 2014). Third, due to its low computational cost compared to microscopic flow simulators,
the developed model may be integrated in a dynamic demand estimator (Hänseler et al., 2014), or employed for realtime
crowd control.