nloptr_fast_options {tsgarch}R Documentation

Default options for nloptr solver

Description

Default options for nloptr solver

Usage

nloptr_fast_options(
  trace = FALSE,
  xtol_rel = 1e-14,
  maxeval = 1000,
  xtol_abs = 1e-12
)

nloptr_global_options(
  trace = FALSE,
  xtol_rel = 1e-14,
  maxeval = 1000,
  xtol_abs = 1e-12
)

Arguments

trace

equivalent to option “print_level” for nloptr. High values results in more details.

xtol_rel

Stop when an optimization step (or an estimate of the optimum) changes every parameter by less than xtol_rel multiplied by the absolute value of the parameter.

maxeval

top when the number of function evaluations exceeds maxeval. This is not a strict maximum: the number of function evaluations may exceed maxeval slightly, depending upon the algorithm

xtol_abs

is a vector of length n (the number of elements in x) giving the tolerances: stop when an optimization step (or an estimate of the optimum) changes every parameter x(i) by less than xtol_abs(i).

Details

These as just a set of pre-created defaults which work well, particularly the “nloptr_fast_options” which uses an SQP solver. nloptr has many other solvers and combination of solvers which can be used. However, keep in mind that the solver must accept analytic derivatives as well as nonlinear inequality constraints.

Value

A list with options which can be passed to the solver.


[Package tsgarch version 1.0.2 Index]