Microscopic models and simulation

Abstract

The movement patterns of pedestrian crowds at train stations, large-scale events, and cities are intricate. Due to the complex interplay between pedestrians, the physical environment, the physiological environment, and the available information, predicting how a pedestrian crowd will move is challenging. Microscopic pedestrian simulation models can help identify which movement patterns will arise, when and where, and to what extent the predicted movement patterns might potentially be problematic. Various pedestrian simulation models have been introduced to model pedestrians’ operational movement dynamics. This chapter explores only one of these simulation model categories: microscopic pedestrian simulation models. We introduce a wide variety of microscopic model types currently used to model pedestrian movement dynamics at a microscopic level, including the Cellular Automata model (CA), the Social Force model (SF), the predictive collision avoidance model (PCA), the Optimal Step Model (OSM), the Discrete Choice model (DC), and the data-driven artificial neural network model (ANN). The overall movement dynamics of each agent result from the interactions between agents, their physical and physiological environment, and available information. At the same time, each model type simulates the movement dynamics of agents in a particular way, resulting in slightly different operational and global movement dynamics. This chapter presents the most naive version of each model type, as this best describes the agents’ general choice behavior and movement dynamics. In addition, some interesting extensions, benefits, and challenges of each model type are discussed.