Abstract:It aims to solve the problem that the evacuation state of pedestrians depicted by the traditional social force model in a crowded multiexit scenario has a relatively large difference with the actual state, especially the 'optimal path' considered by the self-driving force is the problem of shortest path, and the multiexit evacuation mode depicted by the 'herd behavior' is the local optimum problem. Through in-depth analysis of actual evacuation data of pedestrians and causes of problem, a new crowd evacuation optimization model is established in order to effectively improve the simulation accuracy of crowd evacuation in a multi-exit environment. The model obtains the direction of motion of pedestrians using a field model, fully considers the factors such as exit distance, distribution of pedestrians and regional crowding degree, makes a global optimization for the self-driving force in the social force model using a centralized and distributed network model, and makes a local optimization for it using an elephant herding algorithm, so as to establish a new evacuation optimization method for optimal self-adaption in the bottleneck area. The performance status is compared between the improved social force model and the new model by experiments, and the key factors that affect the new model are analyzed in an in-depth manner. The results show that the new model can optimize the optimal path choice at the early stage of evacuation and improve the evacuation efficiency of pedestrians at the late stage, so as to ensure relatively even distribution of pedestrians at each exit, and also make the simulated evacuation process be more real; and the improvement in overall evacuation efficiency is greater when the number of pedestrians to be evacuated is larger. Therefore, the new model provides a method to solve the phenomenon of disorder in overall pedestrian evacuation due to excessive crowd density during the process of multi-exit evacuation.

An extended queueing model based on vision and morality for crowd evacuation

Pedestrians’ behavior in emergency evacuation: Modeling and simulation*

Modifications to Social Force Model

The study on mechanical model considering optimal self-adaption in the bottleneck area

Modeling Domino Effect along the Queue Using an Improved Social Force Model

Exploring the Behavior of Self-Organized Queuing for Pedestrian Flow Through a Non-Service Bottleneck

Behavior-based Cellular Automaton Model for Pedestrian Dynamics

Visual information based social force model for crowd evacuation

A Mixed Crowd Movement Model Incorporating Chasing Behavior

Modeling and Simulation of Crowd Pre-Evacuation Decision-Making in Complex Traffic Environments

Modeling the Dynamics of Pedestrian Evacuation in a Complex Environment

Order–disorder Phase Transitions in Front of the Exit During Human Crowd Evacuations

A modified social force model for crowd evacuation considering collision predicting behaviors

Entropy-Based Crowd Evacuation Modeling with Seeking Behavior of Social Groups

Modeling the Mutual Anticipation in Human Crowds with Attention Distractions

Modeling of Crowd Evacuation with Assailants Via a Fuzzy Logic Approach

Modeling Crowd Evacuation Via Behavioral Heterogeneity-Based Social Force Model

A Vision-Driven Model Based on Cognitive Heuristics for Simulating Subgroup Behaviors During Evacuation

Improved Social Force Model for Emergency Evacuation Scene with Limited Field of Vision

Simulation study on cooperation behaviors and crowd dynamics in pedestrian evacuation

Modeling of Crowd Dynamics with Emergency Signs Via Modified Social Force Model