RK-Opt: A package for the design of numerical ODE solvers

Abstract

Ordinary and partial differential equations (ODEs and PDEs) are used to model many important phenomena. In most cases, solutions of these models must be approximated by numerical methods. Most of the relevant algorithms fall within a few classes of methods, with the properties of individual methods determined by their coefficients. The choice of appropriate coefficients in the design of methods for specific applications is an important area of research. RK-Opt is a software package for designing numerical ODE solvers with coefficients optimally chosen to provide desired properties. It is available from https://github.com/ketch/RK-Opt, with documentation at http://numerics.kaust.edu.sa/RK-Opt/. The primary focus of the package is on the design of Runge-Kutta methods, but some routines for designing other classes of methods such as multistep Runge-Kutta and general linear methods are also included.