UnirootAll {DescTools}R Documentation

Finds many (all) roots of one equation within an interval


The function UnirootAll searches the interval from lower to upper for several roots (i.e., zero's) of a function f with respect to its first argument.


UnirootAll(f, interval, lower = min(interval), upper = max(interval),
            tol = .Machine$double.eps^0.5, maxiter = 1000, n = 100, ...)



the function for which the root is sought.


a vector containing the end-points of the interval to be searched for the root.


the lower end point of the interval to be searched.


the upper end point of the interval to be searched.


the desired accuracy (convergence tolerance).


the maximum number of iterations.


number of subintervals in which the root is sought.


additional named or unnamed arguments to be passed to f (but beware of partial matching to other arguments).


f will be called as f(x, ...) for a numeric value of x.

Run demo(Jacobandroots) for an example of the use of UnirootAll for steady-state analysis.

See also second example of gradient This example is discussed in the book by Soetaert and Herman (2009).


a vector with the roots found in the interval


This is a verbatim copy from rootSolve::uniroot.all (v. 1.7).


The function calls uniroot, the basic R-function.

It is not guaranteed that all roots will be recovered.

This will depend on n, the number of subintervals in which the interval is divided.

If the function "touches" the X-axis (i.e. the root is a saddle point), then this root will generally not be retrieved. (but chances of this are pretty small).

Whereas unitroot passes values one at a time to the function, UnirootAll passes a vector of values to the function. Therefore f should be written such that it can handle a vector of values. See last example.


Karline Soetaert <karline.soetaert@nioz.nl>

See Also

uniroot for more information about input.


## =======================================================================
##   Mathematical examples
## =======================================================================

# a well-behaved case...
fun <- function (x) cos(2*x)^3

curve(fun(x), 0, 10,main = "UnirootAll")

All <- UnirootAll(fun, c(0, 10))
points(All, y = rep(0, length(All)), pch = 16, cex = 2)

# a difficult case...
f <- function (x) 1/cos(1+x^2)
AA <- UnirootAll(f, c(-5, 5))
curve(f(x), -5, 5, n = 500, main = "UnirootAll")
points(AA, rep(0, length(AA)), col = "red", pch = 16)

f(AA)  # !!!

## =======================================================================
## Vectorisation:
## =======================================================================
# from R-help Digest, Vol 130, Issue 27
# https://stat.ethz.ch/pipermail/r-help/2013-December/364799.html

integrand1 <- function(x) 1/x*dnorm(x)
integrand2 <- function(x) 1/(2*x-50)*dnorm(x)
integrand3 <- function(x, C) 1/(x+C)

res <- function(C) {
  integrate(integrand1, lower = 1, upper = 50)$value +
  integrate(integrand2, lower = 50, upper = 100)$value -
  integrate(integrand3, C = C, lower = 1, upper = 100)$value

# uniroot passes one value at a time to the function, so res can be used as such
uniroot(res, c(1, 1000))

# Need to vectorise the function to use UnirootAll:
res <- Vectorize(res)
UnirootAll(res, c(1,1000))

[Package DescTools version 0.99.51 Index]