Forecasting model for pedestrian distribution under emergency evacuation

Abstract

Pedestrian distribution forecasting on the road network is developed to support the evacuation decision-making. The numbers of evacuees distributed on each road link are stochastic, uncertain and multi-dependent. Therefore, a Gaussian Bayesian networks (GBN) based forecasting model is presented, considering the pedestrian flow characteristics, optimization of evacuation route and evacuation decision-making. In the forecasting model, the route choice probabilities obtained by minimizing evacuation time are applied to correct the regression coefficients of GBN. Finally, an example is provided to illustrate the usefulness of this model. Research shows that this model not only reflects the complexity and dynamics of evacuation process but also performs an accurate forecasting on the time development of the pedestrian distributed in the evacuation space.