mize_init {mize}

Initialize the Optimizer.

Description

Prepares the optimizer for use with a specific function and starting point.

Usage

mize_init(
  opt,
  par,
  fg,
  max_iter = Inf,
  max_fn = Inf,
  max_gr = Inf,
  max_fg = Inf,
  abs_tol = NULL,
  rel_tol = abs_tol,
  grad_tol = NULL,
  ginf_tol = NULL,
  step_tol = NULL
)

Arguments

opt

Optimizer, created by make_mize.

par

Vector of initial values for the function to be optimized over.

fg

Function and gradient list. See the documentation of mize.

max_iter

(Optional). Maximum number of iterations. See the 'Convergence' section of mize for details.

max_fn

(Optional). Maximum number of function evaluations. See the 'Convergence' section of mize for details.

max_gr

(Optional). Maximum number of gradient evaluations. See the 'Convergence' section of mize for details.

max_fg

(Optional). Maximum number of function or gradient evaluations. See the 'Convergence' section of mize for details.

abs_tol

(Optional). Absolute tolerance for comparing two function evaluations. See the 'Convergence' section of mize for details.

rel_tol

(Optional). Relative tolerance for comparing two function evaluations. See the 'Convergence' section of mize for details.

grad_tol

(Optional). Absolute tolerance for the length (l2-norm) of the gradient vector. See the 'Convergence' section of mize for details.

ginf_tol

(Optional). Absolute tolerance for the infinity norm (maximum absolute component) of the gradient vector. See the 'Convergence' section of mize for details.

step_tol

(Optional). Absolute tolerance for the size of the parameter update. See the 'Convergence' section of mize for details.

Details

Should be called after creating an optimizer with make_mize and before beginning any optimization with mize_step. Note that if fg and par are available at the time mize_step is called, they can be passed to that function and initialization will be carried out automatically, avoiding the need to call mize_init.

Optional convergence parameters may also be passed here, for use with check_mize_convergence. They are optional if you do your own convergence checking.

Details of the fg list can be found in the 'Details' section of mize.

Value

Initialized optimizer.

Examples

# Create an optimizer
opt <- make_mize(method = "L-BFGS")

# Function to optimize and starting point defined after creating optimizer
rosenbrock_fg <- list(
  fn = function(x) {
    100 * (x[2] - x[1] * x[1])^2 + (1 - x[1])^2
  },
  gr = function(x) {
    c(
      -400 * x[1] * (x[2] - x[1] * x[1]) - 2 * (1 - x[1]),
      200 * (x[2] - x[1] * x[1])
    )
  }
)
rb0 <- c(-1.2, 1)

# Initialize with function and starting point before commencing optimization
opt <- mize_init(opt, rb0, rosebrock_fg)

# Finally, can commence the optimization loop
par <- rb0
for (iter in 1:3) {
  res <- mize_step(opt, par, rosenbrock_fg)
  par <- res$par
  opt <- res$opt
}