Optimized leaders strategies for crowd evacuation in unknown environments with multiple exits

by   Giacomo Albi, et al.

In this chapter, we discuss the mathematical modeling of egressing pedestrians in an unknown environment with multiple exits. We investigate different control problems to enhance the evacuation time of a crowd of agents, by few informed individuals, named leaders. Leaders are not recognizable as such and consist of two groups: a set of unaware leaders moving selfishly toward a fixed target, whereas the rest is coordinated to improve the evacuation time introducing different performance measures. Follower-leader dynamics is initially described microscopically by an agent-based model, subsequently a mean-field type model is introduced to approximate the large crowd of followers. The mesoscopic scale is efficiently solved by a class of numerical schemes based on direct simulation Monte-Carlo methods. Optimization of leader strategies is performed by a modified compass search method in the spirit of metaheuristic approaches. Finally, several virtual experiments are studied for various control settings and environments.



page 14

page 15

page 17

page 18

page 20

page 22

page 23

page 25


Multi Type Mean Field Reinforcement Learning

Mean field theory provides an effective way of scaling multiagent reinfo...

Using Reinforcement Learning to Herd a Robotic Swarm to a Target Distribution

In this paper, we present a reinforcement learning approach to designing...

Hybrid control for optimal visiting problems for a single player and for a crowd

In an optimal visiting problem, we want to control a trajectory that has...

Path Planning in Dynamic Environments using Generative RNNs and Monte Carlo Tree Search

State of the art methods for robotic path planning in dynamic environmen...

Mean-field control variate methods for kinetic equations with uncertainties and applications to socio-economic sciences

In this paper, we extend a recently introduced multi-fidelity control va...

Simulation of Wheelchair Movements in Crowd Using Fine Grid Cellular Automata

Crowd simulation models are used to assess the performance and safety of...
This week in AI

Get the week's most popular data science and artificial intelligence research sent straight to your inbox every Saturday.