{"id":"arxiv/2609.24871","url":"https://arxiv.org/abs/2609.24871","tags":["math.NA"],"score":0,"title":"GradAgent: A Knowledge-Guided Multi-Agent System for Structure-Preserving Gradient-Flow Computation with an Application to Multicomponent Vesicle Dynamics","domain":"math.NA","source":"arxiv","authors":["Zhenlin Guo","Jiale Meng","Shuqi Tang","Haiyan Su","Maosheng Jiang","Kaiwen Shi","Meng Zhao"],"reviews":0,"subject":"math.NA","summary":"High-order differential operators and nonlinear coupling make it challenging to construct conservative and energy-stable schemes for coupled gradient-flow systems. We present GradAgent, a knowledge-guided multi-agent system that coordinates three agents across model analysis, algorithm design and proofs, and numerical implementation and validation. Independent audits strengthen reliability by uncovering mathematical errors and proof gaps, guiding revisions, and maintaining consistency across stages. In Reconstruction Mode, GradAgent reconstructs 20 published studies and organizes audited…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-09-21","reactions":{}}