optim_prox_control: Control parameters for proximal gradient descent
In hoangtt1989/BICEP: Bi-linear algorithm for composite empirical likelihood

Control parameters for proximal gradient descent

1
2
3

optim_prox_control(tol_type = c("par", "fval"), step_size = 1,
  accelerate = F, line_search = T, ls_beta = 0.75, ls_eps = 1e-10,
  ls_maxit = 50, maxit = 1000, eps = 1e-06)

`tol_type`	the type of tolerance checking.
`step_size`	initial step size.
`accelerate`	a boolean to turn on acceleration.
`line_search`	a boolean to turn on line search - it is NOT recommended to turn this off.
`ls_beta`	by how much should the step-size increase or decrease in each line search iteration?
`ls_eps`	you can miss the line search check by this much.
`ls_maxit`	the maximum number of line search iterations.
`maxit`	the maximum number of iterations.
`eps`	the maximum value of the tolerance before declaring convergence.