gradientproj: Constrained optimization with Gradient Projection

In brunasqz/NonlinearOpMethods: Nonlinear Optimization Methods

Description

gradientproj solves a multivariate function with constraints using the Gradient Projection.

Usage

gradientproj(obj.list, x.list, constraint, maxNI = 50, eps = 1e-04,
  alpha0 = 1, c = 1e-04, rho = 0.5, ...)

Arguments

obj.list	Either an objective function or a list with the following names f: objective function df: gradient of f (it not provided, defaults to the numerical version)
x.list	Either a vector with an initial solution or a list with the following names x: a vector with its value in the search space fx: a scalar with its objective value dfx: a vector with its gradient value
constraint	A list, with the following names xmin: lower restriction xmax: upper restraint
maxNI	maximum number of iterations
eps	tolerance for stop codition
alpha0	Initial step size in the backtracking
c	A small constant, control parameter.
rho	A constant to reduce alpha in backtracking
...	parameters used in the check_parameters function

Value

Returns a list with the (approximate) optimum.

References

1. Nocedal, C.T.; Iterative Methods for optimization.

Examples

f <- function(x) {
 f1 <- x[1]^2 + 2*x[2]^2 + 3*x[3]^2 - 2*x[1] -4*x[2] -6*x[3] + 6
 return(f1)
}
x0 <- list(x = c(2.8,3,3.5), fx = f(c(2.8,3,3.5)))
const <- list(xmin = c(2,2,2), xmax = c(4,4,4))
gradientproj(f, x0, const)

#can also be used to list:
obj <- list(functionObj = f)
gradientproj(obj, x0, const)

brunasqz/NonlinearOpMethods documentation built on Oct. 27, 2019, 5:46 a.m.