[16p-K505-1~17] 23.1 Joint Session N "Informatics"

4:30 PM - 4:45 PM JST(7:30 AM - 7:45 AM UTC)Presentation by Applicant for JSAP Young Scientists Presentation Award

[16p-K505-13] Inverse Design of Plastics via Evolutionary Algorithm and Large Language Model

〇Shun Muroga¹, Naoyuki Matsumoto¹, Don N. Futaba¹, Kenji Hata¹ (1.AIST)

Keywords:

Inverse design,multi-objective optimization,large language model

This study proposes a novel approach that combines a large language model, an evolutionary algorithm, and pre-trained predictive models to rapidly suggest materials and process conditions tailored to specific applications and requirements. The method employs plastics as a model system to demonstrate its capability to input requirements for applications such as bumpers or exterior components, perform multi-objective optimization, and provide explanations for differences among candidate Pareto solutions. This approach is expected to accelerate research and development through applications in both closed-loop autonomous experimentation and human-in-the-loop frameworks.