Verification of Fuzzy Rule Interpretation Capabilities in Large Language Models

Abstract

This study aims to verify the fuzzy rule interpretation capabilities of large language models by addressing the optimization problem of battery energy storage systems (BESS) and creating optimal charge and discharge schedules according to given conditions and fuzzy rules. Additionally, we automatically generated fuzzy rules by large language model based on the given conditions, and verified how the generated fuzzy rules and ambiguous expressions are interpreted, as well as how they can be reflected in solving the optimization problem.