## CRAN Task View: Differential Equations

 Maintainer: Karline Soetaert and Thomas Petzoldt Contact: karline.soetaert at nioz.nl Version: 2014-12-26

Differential equations (DE) are mathematical equations that describe how a quantity changes as a function of one or several (independent) variables, often time or space. Differential equations play an important role in biology, chemistry, physics, engineering, economy and other disciplines.

Differential equations can be separated into stochastic versus deterministic DEs. Problems can be split into initial value problems versus boundary value problems. One also distinguishes ordinary differential equations from partial differential equations, differential algebraic equations and delay differential equations. All these types of DEs can be solved in R. DE problems can be classified to be either stiff or nonstiff; the former type of problems are much more difficult to solve.

The dynamic models SIG is a suitable mailing list for discussing the use of R for solving differential equation and other dynamic models such as individual-based or agent-based models.

This task view was created to provide an overview on the topic. If we forgot something, or if a new package should be mentioned here, please let us know.

Stochastic Differential Equations (SDEs)

In a stochastic differential equation, the unknown quantity is a stochastic process.
• The package sde provides functions for simulation and inference for stochastic differential equations. It is the accompanying package to the book by Iacus (2008).
• The package pomp contains functions for statistical inference for partially observed Markov processes.
• Packages adaptivetau and GillespieSSA implement Gillespie's "exact" stochastic simulation algorithm (direct method) and several approximate methods.

Ordinary Differential Equations (ODEs)

In an ODE, the unknown quantity is a function of a single independent variable. Several packages offer to solve ODEs.
• The "odesolve" package was the first to solve ordinary differential equations in R. It contains two integration methods. It is not actively maintained and has been replaced by the package deSolve.
• The package deSolve contains several solvers for solving ODEs. It can deal with stiff and nonstiff problems.
• The package deTestSet contains solvers designed for solving very stiff equations.

Delay Differential Equations (DDEs)

In a DDE, the derivative at a certain time is a function of the variable value at a previous time.
• The package PBSddesolve (originally published as "ddesolve") includes a solver for non-stiff DDE problems.
• Functions in the package deSolve can solve both stiff and non-stiff DDE problems.

Partial Differential Equations (PDEs)

PDEs are differential equations in which the unknown quantity is a function of multiple independent variables. A common classification is into elliptic (time-independent), hyperbolic (time-dependent and wavelike), and parabolic (time-dependent and diffusive) equations. One way to solve them is to rewrite the PDEs as a set of coupled ODEs, and then use an efficient solver.
• The R-package ReacTran provides functions for converting the PDEs into a set of ODEs. Its main target is in the field of ''reactive transport'' modelling, but it can be used to solve PDEs of the three main types. It provides functions for discretising PDEs on cartesian, polar, cylindrical and spherical grids.
• The package deSolve contains dedicated solvers for 1-D, 2-D and 3-D time-varying ODE problems as generated from PDEs (e.g. by ReacTran).
• Solvers for 1-D time varying problems can also be found in the package deTestSet.
• The package rootSolve contains optimized solvers for 1-D, 2-D and 3-D algebraic problems generated from (time-invariant) PDEs. It can thus be used for solving elliptic equations.
Note that, to date, PDEs in R can only be solved using finite differences. At some point, we hope that finite element and spectral methods will become available.

Differential Algebraic Equations (DAEs)

Differential algebraic equations comprise both differential and algebraic terms. An important feature of a DAE is its differentiation index; the higher this index, the more difficult to solve the DAE.
• The package deSolve provides two solvers, that can handle DAEs up to index 3.
• Three more DAE solvers are in the package deTestSet.

Boundary Value Problems (BVPs)

BVPs have solutions and/or derivative conditions specified at the boundaries of the independent variable.
• Package bvpSolve deals only with this type of equations.
• The package ReacTran can solve BVPs that belong to the class of reactive transport equations.

Other

• The simecol package provides an interactive environment to implement and simulate dynamic models. Next to DE models, it also provides functions for grid-oriented, individual-based, and particle diffusion models.
• Packages scaRabee and PKfit offer frameworks for simulation and optimization of Pharmacokinetic-Pharmacodynamic Models.
• In the package FME are functions for inverse modelling (fitting to data), sensitivity analysis, identifiability and Monte Carlo Analysis of DE models.
• The packages nlmeODE and PSM have functions for mixed-effects modelling using differential equations.
• mkin provides routines for fitting kinetic models with one or more state variables to chemical degradation data.
• The package CollocInfer implements collocation-inference for continuous-time and discrete-time stochastic processes.
• Root finding, equilibrium and steady-state analysis of ODEs can be done with the package rootSolve.
• The deTestSet package contains many test problems for differential equations.
• Package pracma contains solvers for ODEs, as pure R scripts, useful as a learning tool.
• The PBSmodelling package adds GUI functions to models.
• Package ecolMod contains the figures, data sets and examples from a book on ecological modelling (Soetaert and Herman, 2009).
• primer is a support package for the book of Stevens (2009).