Multi-agent reinforcement learning using echo-state network and its application to pedestrian dynamics

Abstract

In recent years, simulations of pedestrians using the multi-agent reinforcement learning (MARL) have been studied. This study considered the roads on a grid-world environment, and implemented pedestrians as MARL agents using an echo-state network and the least squares policy iteration method. Under this environment, the ability of these agents to learn to move forward by avoiding other agents was investigated. Specifically, we considered two types of tasks: the choice between a narrow direct route and a broad detour, and the bidirectional pedestrian flow in a corridor. The simulations results indicated that the learning was successful when the density of the agents was not that high.

arXiv comment25 pages, 19 figures

References in corpus (9)

• Novel type of phase transition in a system of self-driven particles
• Social Force Model for Pedestrian Dynamics
• Jamming at Zero Temperature and Zero Applied Stress: the Epitome of Disorder
• The Fundamental Diagram of Pedestrian Movement Revisited
• The Jamming Transition in Granular Systems
• Active Jamming: Self-propelled soft particles at high density
• Distributive Dynamic Spectrum Access through Deep Reinforcement Learning: A Reservoir Computing Based Approach
• Experimental study of pedestrian counterflow in a corridor
• A Survey on Reservoir Computing and its Interdisciplinary Applications Beyond Traditional Machine Learning