Tutorial: Simulating Differential Equation Models in R
Karline
Soetaert, Netherlands Institute for Ecology,
Netherlands.
Thomas
Petzoldt, Dresden University of Technology,
Germany.
Abstract
R has become the most widely used systems for statistical data
analysis, but it is also well suited for other disciplines in
scientific computing. One of the fields where considerable
progress has been made is the solution of differential
equations. Package deSolve provides the useR with
state-of-the-art technology for handling differential equations
in R. The tutorial will concentrate on solving initial value
problems of ordinary differential equations (ODE). We will give
an overview over the different solver functions available in
packages deSolve and rootSolve. Practical examples will show
that even numerically challenging systems can be efficiently
solved in R and how external data can be handled. An outlook
will demonstrate how partial differential equations (PDE) for
reaction diffusion systems in 1D, 2D or 3D can be handled in R
and how impressive computation performance can be approached.
Outline
-
How to specify a model: Differential
equation modelling made easy.
-
Don't be afraid of stiffness: An overview
over the solver functions.
-
Dynamics, chaos and equilibria: Plotting,
scenario comparison and root finding.
-
Under control: Forcing functions and
events.
-
Diffusion, advection and reaction: Partial
differential equations (PDE) with ReacTran.
-
R without handbrake: Using matrices and
compiled code.
Prerequisites
It will be assumed that participants have good knowledge of
the R language and are familiar with ordinary differential
equation models in any discipline. At least basic knowledge
about numerical simulation methods would be helpful.
Preparatory material can be found on the deSolve web page.
Intended Audience
Differential equations can be used to describe exchanges of
matter, energy, information or any other quantities as they
vary in time and/or space.
We invite people with background in natural, environmental
and life sciences, as well as systems sciences, engineering,
economics or any other discipline to experience how nicely
and flexibly R can be used to explore time-dependend behavior
of their dynamical systems.
Pre-conference discussion is possible on the mailig list: r-sig-dynamic-models@r-project.org
Please check here for up to date tutorial resources.