Spatial point process simulator

Spatial point process simulator based on accept/reject algorithm.

Usage

acceptreject(n, f = NULL, x0 = 0, x1 = 1, y0 = 0, y1 = 1,
  m = 0, maxit = 100, fail = FALSE)

Arguments

n

number of points to generate.

f

a function returning probability (value between 0 and 1) given distance as the first and only argument. The function generates spatially uniform Poisson point process (complete spatial randomness) when NULL.

x0, x1, y0, y1

x and y ranges (bounding box).

m

margin width for avoiding edge effects.

maxit

maximum number of iterations per point to try if no acceptance happens.

fail

logical, what to do when there is a problem. TRUE gives error, the default FALSE gives only a warning.

Value

A matrix with n rows and 2 columns for x and y coordinates.

Author

Peter Solymos

Examples

## complete spatial randomness
plot(acceptreject(100), asp=1)


## more systematic
distance <- seq(0,1,0.01)
f <- function(d)
  (1-exp(-d^2/0.1^2) + dlnorm(d, 0.2)/dlnorm(exp(0.2-1),0.2)) / 2
op <- par(mfrow = c(1, 2))
plot(distance, f(distance), type="l")
plot(acceptreject(100, f, m=1), asp=1)

par(op)

## more clustered
f <- function(d)
  exp(-d^2/0.1^2) + 0.5*(1-exp(-d^2/0.4^2))
op <- par(mfrow = c(1, 2))
plot(distance, f(distance), type="l")
plot(acceptreject(100, f, m=1), asp=1)

par(op)