An electrical engineering perspective on missed opportunities in computational physics

Abstract

We look at computational physics from an electrical engineering perspective and suggest that several concepts of mathematics, not so well-established in computational physics literature, present themselves as opportunities in the field. We emphasize the virtues of the concept of elliptic complex and highlight the category theoretical background and its role as a unifying language between algebraic topology, differential geometry and modelling software design. In particular, the ubiquitous concept of naturality is central. We discuss the Galerkin finite element method as a way to achieve a discrete formulation and discuss its compatibility with so-called cochain methods. Despite the apparent differences in their underlying principles, in both one finds a finite-dimensional subcomplex of a cochain complex. From such a viewpoint, compatibility of a discretization boils down to preserving properties in such a process. Via reflection on the historical background and the identification of common structures, forward-looking research questions may be framed.