Operator-splitting (OS) methods are powerful and popular divide-and-conquer methods used to solve differential equations. Implementations of OS methods, however, are rarely performed in a general-purpose way. This paper introduces pythOS, a Python library designed to facilitate the systematic implementation, exploration, and evaluation of OS methods. Two examples are presented to demonstrate the library’s functionality and versatility: a model of neutron transport in a high-scattering medium and the monodomain model of cardiac electrophysiology. Both examples employ different OS methods and numerical sub-integration techniques to demonstrate some of pythOS’s capabilities. The results show the utility of pythOS as a general-purpose tool for researchers and practitioners for the time integration of differential equations using OS methods.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

pythOS: An Operator-Splitting Library in Python

  • Mohammad Mahdi Moayeri,
  • Victoria Guenter,
  • Siqi Wei,
  • Raymond J. Spiteri

摘要

Operator-splitting (OS) methods are powerful and popular divide-and-conquer methods used to solve differential equations. Implementations of OS methods, however, are rarely performed in a general-purpose way. This paper introduces pythOS, a Python library designed to facilitate the systematic implementation, exploration, and evaluation of OS methods. Two examples are presented to demonstrate the library’s functionality and versatility: a model of neutron transport in a high-scattering medium and the monodomain model of cardiac electrophysiology. Both examples employ different OS methods and numerical sub-integration techniques to demonstrate some of pythOS’s capabilities. The results show the utility of pythOS as a general-purpose tool for researchers and practitioners for the time integration of differential equations using OS methods.