Solvers for Initial Value Problems of Differential Equations ('ODE', 'DAE', 'DDE')

Functions that solve initial value problems of a system of first-order ordinary differential equations ('ODE'), of partial differential equations ('PDE'), of differential algebraic equations ('DAE'), and of delay differential equations. The functions provide an interface to the FORTRAN functions 'lsoda', 'lsodar', 'lsode', 'lsodes' of the 'ODEPACK' collection, to the FORTRAN functions 'dvode', 'zvode' and 'daspk' and a C-implementation of solvers of the 'Runge-Kutta' family with fixed or variable time steps. The package contains routines designed for solving 'ODEs' resulting from 1-D, 2-D and 3-D partial differential equations ('PDE') that have been converted to 'ODEs' by numerical differencing.


News

Changes version 1.21

o change way how PROTECT/UNPROTECT is handled (Thomas) (suggested by Tomas Kalibera / R-Core) o fixed inconsistency in the aphid model example (suggested by Sarah Kintner)

Changes version 1.20

o register native routines (Thomas) o check if event data frame has ordered time (and if not, order) o change 'event list' to event matrix or data frame in docs o intentional version jump to indicate chances at the C level

Changes version 1.14

o matplot.deSolve is not anymore exported as matplot to avoid the respective startup message o please use matplot.deSolve or the alias matplot.0D instead (Thomas) o small fix that allows parameters in list format for DLLfunc and DLLres o a little bit Fortran modification (e.g. avoid real8 and complex16 types)

Changes version 1.13

o observed data and plot.deSolve / matplot for multiple outputs (Karline) o combining compiled code function with R code event function (Karline) o check sorting of event times (Karline) o fix bug related to negative event time (patch supplied by J. Stott) o relax setting of tcrit to make integration with events slightly faster (patch from J. Stott) o adapt maxstep calculation for rk methods, print a warning if maxsteps is exceeded, fix diagnostics (Thomas) o more argument checking for rk solvers (Thomas) o add reference to book of Soetaert, Cash and Mazzia (2012)

Changes version 1.12

o new functions matplot.deSolve and matplot.1D o fix valgrind issue (detected by new compilers) o small improvments of plotting functions o import standard packages as required by upcoming R versions

Changes version 1.11

o compiledCode vignette now with dede example o warning and error bug resolved o Time SEXP incompatibility with R 3.1.1 resolved o CFunc compatibility (compiled code)

Changes version 1.10.9

o documentation updates, hyperlinks to examples and vignettes o moved example directories

Changes version 1.10.8

o remove redundant .R files from inst/doc o fixed bug in event code (patch contributed by Jonathan Stott)

Changes version 1.10.7

o Fortran examples of compiled dede models (Woody) o vignettes moved to /vignettes o roles of authors ([email protected]) o function timestep is now internal o small documentation updates

Changes version 1.10.6 (Thomas)

o change declaration of variable dimensions from (1) to (*) in legacy Fortran code to pass automatic bounds check o remove the Jacobian examples from ?ode because banddown=0 can lead to problems on some systems; examples will come back in a next release o fixed bug in the "iteration" solver o small documentation updates

Changes version 1.10.5 (Karline, Thomas)

o extended subset.deSolve with argument arr, when TRUE returns an array for >2-D output o fixed the R compiler notes o plot.ode.2D now has an mtext argument, via the ..., to label multiple figures in margin... CHECK - see ode.2D o subset can also be a vector with indices in addition to logical o image with legend = TRUE changed size of plot in different layouts - now solved (by adding par(mar = par("mar")) ) o new method to output warnings and error messages o add data type check for external outputs in rk_util.c o add interface for compiled dede models o emphasize consistent order of states in y and return value of func o changes of Fortran error messages (to be continued)

Changes version 1.10-4 (Thomas, Karline)

o allow reverted time vector for fixed step solvers

  • todo: find solution for dense output methods, and Livermore solvers o all solvers now have default atol = 1e-6; before this daspk and vode had 1e-8. o multiple warnings from daspk if num steps = 500 toggled off. o added input argument "nind" to daspk, to make it compatible with radau. this also changes the way the variables are weighed, hence this differs from the original daspk 2.0 code. o improved warning printing in daspk and vode o extended sparse Jacobian input in lsodes. (2-D and 3-D sparsity with mapping var and arbitrary sparsity in ian/jan format).

Changes version 1.10-3 (Karline)

o rwork and iwork in lsodes from Fortran -> C (to remove compiler warnings) o roots + events: now certain roots can stop simulation + fixed bug in radau root o improved events\roots help file o diagnostics(out) gave error in case method=iteration (no rstate) now fixed o the package authors agreed to assign the maintainer role to T.P., but the order of authorship and credits remain unchanged.

Changes version 1.10-2 (Karline)

o remove NAs from forcing functions - when used in DLL (file forcings.R) o new argument "restructure" in ode.1D, for use with implicit solvers not in deSolve o removed requirement to have eventfunc in compiled code when func is in compiled code o subsetting on summary.deSolve

Changes version 1.10-1 (Thomas)

o remove several redundant variables from C code o add NEWS file

Changes version 1.10 (Karline, Thomas)

o compiled code using mass in daspk o cleanEventTimes

Changes version 1.9+ (Karline)

o roots, events, lags in radau o roots in lsodes o lags in daspk o ode (method = "iteration")

Changes version 1.9 (Karline, Thomas)

o summary.deSolve o subset.deSolve o plotting deSolve objects improved:

  • plot more than one output in same figures (scenarios),
  • add observations o vignette improved o fixed bug in 'timesteps'

Changes version 1.8.1 (Thomas, Woody, Karline)

o fixed compiler warnings using valgrind o fixed compiler warning C-code

Changes version 1.8 (Thomas)

o Dormand-Prince 8(7) coefficients use now common instead of decimal fractions

Changes version 1.8 (Karline)

o Runge-Kuttas:

  • extra output: number of failed steps (see also 2)
  • number of function evaluations + 1 for initial condition
  • dense output for cash-karp
  • dopri8(7) added
  • radau added!! implicit runge kutta, solves also DAE up to index 3! o other:
  • image function for ode.2-D added.
  • changed warning printing in FORTRAN code
  • common interface for radau and daspk: both can solve systems written as M*dy = f(x,y). daspk can also solve systems written as 0 = g(x,y,dy) (=default for daspk)

Reference manual

It appears you don't have a PDF plugin for this browser. You can click here to download the reference manual.