pgOptim: Proximal gradient optimisation for non smooth functions

Description Usage Arguments Value

View source: R/optim.R

Description

Function for minimising piecewise differentiable objective functions.

Usage

1
pgOptim(theta, fn, df, prox, acl = F, tol = 1e-06, t)

Arguments

theta

Initial value.

fn

Function to optimise.

df

Subgradient of the objective function.

prox

Proximal operator for proximal gradient method.

acl

Flag for accelerated proximal gradient method. Default is false.

tol

Relative convergence tolerance

t

Stepsizes for updating the set of parameters.

Value

The function returns a list with following components

par

The best set of parameters

value

The value of the function corresponding to best set of parameters.

iteration

Number of iterations taken


tathagatabasu/SubgradOptim documentation built on Nov. 5, 2019, 10 a.m.