Reward Function Design for Crowd Simulation via Reinforcement Learning

TL;DR

This paper investigates how to design reward functions for reinforcement learning-based crowd simulation, providing theoretical insights and empirical evaluations to improve the realism and efficiency of virtual crowds.

Contribution

It offers a theoretical analysis of reward function validity and empirically evaluates energy-based rewards with guiding potentials for crowd simulation.

Findings

Direct energy minimization is effective with proper scaling.

Reward components significantly influence crowd behavior.

Guiding potentials improve energy efficiency in simulations.

Abstract

Crowd simulation is important for video-games design, since it enables to populate virtual worlds with autonomous avatars that navigate in a human-like manner. Reinforcement learning has shown great potential in simulating virtual crowds, but the design of the reward function is critical to achieving effective and efficient results. In this work, we explore the design of reward functions for reinforcement learning-based crowd simulation. We provide theoretical insights on the validity of certain reward functions according to their analytical properties, and evaluate them empirically using a range of scenarios, using the energy efficiency as the metric. Our experiments show that directly minimizing the energy usage is a viable strategy as long as it is paired with an appropriately scaled guiding potential, and enable us to study the impact of the different reward components on the behavior of the simulated crowd. Our findings can inform the development of new crowd simulation techniques, and contribute to the wider study of human-like navigation.