optimizeWithTrace {spatstat.utils}

One Dimensional Optimization with Tracing

Description

Find the minimum or maximum of a function over an interval of real numbers, keeping track of the function arguments and function values that were evaluated.

Usage

optimizeWithTrace(f, interval, ...,
                  lower = min(interval), upper = max(interval))

Arguments

f

The function to be minimized or maximized.

interval

Numeric vector of length 2 containing the end-points of the interval to be searched.

lower, upper

The lower and upper endpoints of the interval to be searched.

...

Other arguments passed to optimize, including arguments to the function f.

Details

This is a simple wrapper for the optimization routine optimize. The function f will be optimized by computing its value at several locations in the interval, as described in the help for optimize. This wrapper function stores the locations and resulting function values, and returns them along with the result of the optimization.

Value

A list with components

• minimum (or maximum), the location in the search interval which yielded the optimum value;

• objective, the value of the function at this location;

• x, the sequence of locations in the interval that were considered (in the order considered);

• y, the function values corresponding to x.

Author(s)

Adrian Baddeley [e-mail], Rolf Turner [e-mail] and Ege Rubak [e-mail].

See Also

optimize

Examples

  f <- function (x, a) (x - a)^2
  result <- optimizeWithTrace(f, c(0, 1), tol = 0.0001, a = 1/3)
  result
  curve(f(x, 1/3))
  with(result, points(x, y, pch=16))