Abstract: In this paper, we propose a multi-agent task planning framework called Fuzzy-LLM, which leverages large language models to transform highly abstract and ambiguous task instructions into ...