Heterogeneous Multi-Agent Reinforcement Learning based on Modularized Policy Network

Open Webpage

Hyeong Tae Kim, Ju-Youn Park

Published: 2025, Last Modified: 23 Oct 2025Expert Syst. Appl. 2025EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•Design a new modular policy in MARL to effectively handle heterogeneous teams.•The proposed policy modules adapt to unseen environments through indirect training.•Our architecture simultaneously learns multiple tasks and agents without forgetting.