Efficient order-adaptive methods for polymer self-consistent field theory
Efficient order-adaptive methods for polymer self-consistent field theory
复制标题
DOI：
10.1016/j.jcp.2019.02.027
复制
发表时间：
2019-06
期刊：
影响因子：
--
通讯作者：
Héctor D. Ceniceros
中科院分区：
文献类型：
--
作者：
Héctor D. Ceniceros
提出了一种高精度、高存储效率的高分子自洽场理论（SCFT）求解方法。其核心思想是结合联合收割机光谱积分的聚合物链轮廓变量与光谱延迟校正技术，以解决SCFT修改的扩散方程具有任意高的精度。其结果是一个强大的方法，实现高精度与最少数量的离散轮廓节点，这转化为大大减少内存需求和提高计算效率。特别是，这种光谱延迟校正方法使计算强隔离系统具有前所未有的精度。此外，延迟校正的框架允许我们在外鞍点迭代期间自适应地增加精度的阶数，以大幅降低SCFT计算的成本。
A highly accurate and memory-efficient approach for the solution of polymer self-consistent field theory (SCFT) is proposed. The central idea is to combine spectral integration in the polymer chain contour variable with a spectral deferred correction technique to solve the SCFT modified diffusion equations with arbitrarily high order of accuracy. The result is a robust method that achieves high accuracy with a minimal number of discrete contour nodes, which translates into vastly reduced memory requirements and increased computational efficiency. In particular, this spectral deferred correction method enables the computation of strongly segregated systems with unprecedented accuracy. Moreover, the framework of deferred corrections allows us to adaptively increase the order of accuracy during the outer saddle point iteration to drastically reduce the cost of a SCFT computation.