Multi-agent reinforcement learning using echo-state network and its application to pedestrian dynamics
This work addresses pedestrian dynamics simulation for urban planning or crowd management, but it is incremental as it applies existing MARL methods to a specific grid-world scenario.
The study tackled pedestrian simulation using multi-agent reinforcement learning (MARL) with echo-state networks and least squares policy iteration, and found that agents successfully learned to move forward by avoiding others in tasks like route choice and bidirectional flow, but only when agent density was not too high.
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.