This paper examines a novel type of multi-agent problem, in which an agent makes multiple identical copies of itself in order to achieve a single agent task better or more efficiently. This strategy improves performance if the environment is noisy and the task is sometimes unachievable by a single agent copy. We propose a learning algorithm for this multicopy problem which takes advantage of the structure of the value function to efficiently learn how to balance the advantages and costs of adding additional copies.
@article{arxiv.2309.10908,
title = {Multicopy Reinforcement Learning Agents},
author = {Alicia P. Wolfe and Oliver Diamond and Brigitte Goeler-Slough and Remi Feuerman and Magdalena Kisielinska and Victoria Manfredi},
journal= {arXiv preprint arXiv:2309.10908},
year = {2025}
}