Learn R Programming

optextras (version 2019-12.4)

Tools to Support Optimization Possibly with Bounds and Masks

Description

Tools to assist in safely applying user generated objective and derivative function to optimization programs. These are primarily function minimization methods with at most bounds and masks on the parameters. Provides a way to check the basic computation of objective functions that the user provides, along with proposed gradient and Hessian functions, as well as to wrap such functions to avoid failures when inadmissible parameters are provided. Check bounds and masks. Check scaling or optimality conditions. Perform an axial search to seek lower points on the objective function surface. Includes forward, central and backward gradient approximation codes.

Copy Link

Version

Install

install.packages('optextras')

Monthly Downloads

289

Version

2019-12.4

License

GPL-2

Maintainer

Last Published

December 20th, 2019

Functions in optextras (2019-12.4)

grback

Backward difference numerical gradient approximation.
ctrldefault

set control defaults
grcentral

Central difference numerical gradient approximation.
grchk

Run tests, where possible, on user objective function and (optionally) gradient and hessian
gHgen

Generate gradient and Hessian for a function at given parameters.
bmstep

Compute the maximum step along a search direction.
gHgenb

Generate gradient and Hessian for a function at given parameters.
axsearch

Perform axial search around a supposed minimum and provide diagnostics
fnchk

Run tests, where possible, on user objective function
hesschk

Run tests, where possible, on user objective function and (optionally) gradient and hessian
optextras-package

Tools to Support Optimization Possibly with Bounds and Masks
scalechk

Check the scale of the initial parameters and bounds input to an optimization code used in nonlinear optimization
kktchk

Check Kuhn Karush Tucker conditions for a supposed function minimum
grnd

A reorganization of the call to numDeriv grad() function.
grfwd

Forward difference numerical gradient approximation.
bmchk

Check bounds and masks for parameter constraints used in nonlinear optimization