ddesolve: DDE solver in R research :: home  

NOTE: The package ddesolve is now deprecated; its successor is PBSddesolve.

PBSddesolve (previously known as ddesolve) is an R port of Solv95 by Alex Couture-Beil, Jon Schnute and Rowan Haigh from Fisheries and Oceans Canada's Pacific Biological Station. A few years ago I contributed some code to extend ddesolve's capabilities.

PyDDE is a Python port of Solv95 that used the same back-end as ddesolve. I assume that similar code is still present in PBSddesolve, but no further development on PyDDE is taking place.

From the users' guide:

PBSddesolve generates numerical solutions for systems of delay differential equations (DDEs) and ordinary differential equations (ODEs). The numerical routines come from Simon Wood's program solve95, originally written in C for the Microsoft Windows operating systems. With PBSddesolve, a user can write the gradient code for a system of DDEs or ODEs in the R language, rather than C. The code will then run on all platforms supported by PBSddesolve, and the results can be inspected using R's extensive graphics capabilities. Simon has generously given us permission to publish PBSddesolve (including his embedded routines) under the GNU GENERAL PUBLIC LICENSE Version 2.

Download. You can obtain PBSddesolve (and even the deprecated ddesolve) from CRAN.

My contribution. Unlike Solv95, since version 0.57 ddesolve has supported output at times specified by a numeric vector. Values at these times are computed during integration using the same cubic Hermite interpolation routines used to compute delayed state variables, and hence have the same degree of accuracy.

Solution to a Lotka-Volterra predator-prey model.
 
 
Solution to Gurney & Nisbet's (1981) model for Nicholson's (1954) blowflies.
 
 
Solution to a DDE with switches.
 
© 2007 B.J. Cairns.
© 2002-2010 Benjamin J. Cairns: e-mail ; ph +44 1865 289673 ;
Cancer Epidemiology Unit, University of Oxford, Richard Doll Building, Roosevelt Drive, Oxford OX3 7LF, U.K.