Network Simulation for Pedestrian Flows with HyDEFS

Abstract

The reliable simulation of pedestrian movement is an essential tool for the security aware design and analysis of buildings and infrastructure. We developed HyDEFS, an event-driven dynamic flow simulation software which is designed to simulate pedestrian movement depending on varying routing decisions of the individual users and varying constraints. HyDEFS uses given density depending velocities to model congestions and evaluates flow distributions with respect to average and maximum travel time. This is of particular importance when considering evacuation scenarios. We apply HyDEFS on two small networks and cross validate its results by time-discrete and time-continuous calculations.

Higher investment levels into pre-planned routes increase the adherence of pedestrians to them

“…is a measure for how busy the edge between node and is. Busyness is measured via the expected speed a pedestrian will have on the edge (Klamroth et al, 2020), as follows:…”

“…The model we used for pedestrian movement on edges is directly based on the previously published model by Klamroth et al (2020). In this model, pedestrian dynamics are driven by events.…”

“…Based on previous work, we used a piece-wise linear function to describe the speed-density relationship (Klamroth et al, 2020). If the current number of pedestrians on an edge is greater than an upper threshold or smaller than a lower threshold value, the speed will be the minimum or maximum speed, respectively; Otherwise, the speed decreases linearly from the maximum speed to the minimum, as described below:…”

“…However, when people are not familiar with the building they are in, they tend to take the movement of others, including congestion, as directional information and may thus prefer the congested route (Tong & Bode, 2021). Several measures for congestion and related factors have been developed, including assessments of the smoothness of pedestrian streams (Feliciani & Nishinari, 2020), the route busyness calculated by the expected speed a pedestrian will have (Klamroth et al, 2020), the crowd pressure (local variance of velocity multiplied by the corresponding density) (Helbing et al, 2007), a congestion number (the ratio of measures with the case of extreme congestion) (Zanlungo et al, 2020), and the estimation of the time lost by passing through congestion (Crociani et al, 2015).…”