pyMOR - Model Order Reduction with Python (PDESoft 2014)

2014-08-11T08:26:39Z (GMT) by Rene Milk
<p>Presentation given at PDESoft 2014 in Heidelberg. Abstract:</p> <p>In this contribution we present pyMOR, a new open source library<br>for building model order reduction applications with the Python programming language.<br>Its main focus lies on the easy application of reduced basis methods<br> to parameterized partial<br>differential equations solved by external high-dimensional discretization packages.<br>At the same time, pyMOR provides a built-in discretization toolkit based on the<br>NumPy/SciPy software stack, which makes it an attractive tool for<br>teaching and rapid exploration of new model reduction algorithms.</p> <p>We give a brief overview of pyMOR's main components and explain in detail the<br>design philosophy and choices that allow us to arrive at a set of completely generic<br>reduction algorithms which can be readily incorporated with any external solver connected<br>to pyMOR via its lightweight operator and vector interfaces.<br>We discuss how this approach simultaneously empowers the solver with both<br>the flexibility and interactivity of a modern dynamic language like Python.<br>We conclude our presentation with a practical example of the provided infrastructure to<br>integrate existing DUNE discretizations with pyMOR.</p> <p> </p>