Quantitative Comparison of Field-Update Algorithms for Polymer SCFT and FTS
Quantitative Comparison of Field-Update Algorithms for Polymer SCFT and FTS
复制标题
DOI：
10.1021/acs.macromol.1c01804
复制
发表时间：
2021-10
期刊：
影响因子：
5.5
通讯作者：
Daniel L. Vigil;K. Delaney;G. Fredrickson
中科院分区：
文献类型：
--
作者：
Daniel L. Vigil;K. Delaney;G. Fredrickson
分析了聚合物自洽场理论（SCFT）和场理论模拟（FTSs）的各种场更新算法。我们提供了第一个直接比较安德森混合和虚构弛豫动力学的SCFT，并发现当两种方案都适当调整时，几乎等效的性能。我们还表明，尽管每一步成本增加，但预测校正算法在虚拟动态方法中是最有效的。对于FTS，发现自适应时间步进显着提高了非均匀系统的算法稳定性，并且能够在比以前实现的更低的链长和密度下进行模拟。
A wide range of field-update algorithms for polymer self-consistent field theory (SCFT) and field theoretic simulations (FTSs) are analyzed. We provide the first direct comparison between Anderson mixing and fictitious relaxational dynamics for SCFT and find nearly equivalent performance when both schemes are properly tuned. We also show that predictor–corrector algorithms are the most efficient among fictitious dynamics approaches despite increased costs per step. For FTS, adaptive time stepping is found to dramatically improve algorithm stability for inhomogeneous systems and enable simulation at much lower chain length and density than was previously achievable.