Insights from liquid crystal models
for designing numerical methods of gradient flows
报告专家：徐劼 副研究员（中国科学院计算所）
报告时间：4月17日（周五）09:30-10:30
报告地点：国家天元数学西南中心117
报告摘要：
Liquid crystal models, except a few ones such Landau-de Gennes, usually possess very complicated forms, which seem formidable when discussing numerical schemes and their analyses. However, the complicated forms they posess are able to drive us to figure out the most essential structures in designing numerical methods. Actually, the idea of scalar auxiliary variables is inspired by considering liquid crystal models. Furthermore, we may need to reexamine techniques that have proved to be powerful for gradient flows or other energy dissipative systems in simple forms, such as maximum principle. In addition, liquid crystal models could have highly coupled constraints, such as range of tensors or constraints of an orthonormal frame. They lead to interesting and challenging problems for numerical methods, for which we present some useful approaches.
专家简介：
徐劼博士于2010年和2015年在北京大学分别获学士和博士学位，后在普渡大学从事博士后研究，现任职于中国科学院数学与系统科学研究院计算数学与科学工程计算研究所。徐劼博士在液晶建模与模拟、梯度流的数值方法方面取得了突出的原创成果，代表工作包括：针对复杂液晶分子发展了基于分子理论的系统建模方法；独立提出了拟熵这一可计算建模的有力工具；提出了梯度流的标量辅助变量方法。获2024 ICCM Distinguished Paper Award 和 International Congress of Basic Science 2025 Frontiers of Science Award。
邀请人：唐庆粦
